Journal of Industrial and Engineering Chemistry 41 (2016) 165–174

Contents lists available at ScienceDirect

Journal of Industrial and Engineering Chemistry

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

Synthesis of rectorite/Fe3O4-CTAB composite for the removal of nitrate

and phosphate from water
Fei Wang a, Dan Liu b, Pengwu Zheng c, Xiaofei Ma a,*
a
Chemistry Department, School of Science, Tianjin University, Tianjin 300072, China
b
School of Physics and Materials Engineering, Dalian Nationalities University, Dalian 116600, China
c
School of Pharmacy, Jiangxi Science and Technology Normal University, Nanchang, Jiangxi 330013, China

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

Article history: The rectorite/Fe3O4-CTAB composite (REC/Fe3O4-CTAB) was fabricated by introducing cetyl trimethyl
Received 24 August 2015 ammonium bromide (CTAB) and Fe3O4 onto the layers of raw rectorite (REC). The layers of rectorite were
Received in revised form 5 March 2016 intercalated or exfoliated, and CTAB enhanced the electrostatic attraction by quaternary ammonium
Accepted 14 July 2016
cations towards nitrate and phosphate anions. The order of Fe3O4 and CTAB introduction had great effect
Available online 25 July 2016
on the adsorption. When Fe3O4 was loaded before CTAB on REC, the obtained composites exhibited
better adsorption. The maximum adsorption capacities could reach 182.1 mg/g for NO3 and 174.5 mg/g
Keywords:
for PO43. Besides, REC/Fe3O4-CTAB composite could be easily regenerated with NaOH solution.
Rectorite
ß 2016 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights
Modification
Nitrate reserved.
Phosphate
Adsorption

Introduction
Agricultural nitrate and phosphate pollution has become the
most predominant source of water pollution in China. Excessive use
of nitrogenous and phosphorous fertilizers causes eutrophication,
which stimulates the overgrowth of algae and aquatic plants,
leading to human health disorders, such as blue-baby syndrome,
infant methemoglobinemia as well as cancer [1]. The present
available methods for the removal of nitrate and phosphate include
ion exchange [2–4], reverse osmosis [5], adsorption [6–13], catalytic
[14,15] and biological denitrification [16]. Among all the methods
studied, adsorption method was attached great importance due to
its high efficiency, simple equipment required and good reliability
[17]. Many natural materials can be good adsorbents for nitrate and
phosphate in water, especially when they are modified, such as
bamboo powder charcoal [18], Fe3O4/ZrO2/chitosan composites
[19], mixed lanthanum/aluminum modified montmorillonite [20],
Fe(III)-montmorillonites [21], positively charged kaolinites [22].
Rectorite (REC) is an interstratified natural clay mineral
composed of a regular stacking of dioctahedral mica-like layer
and dioctahedral smectite-like layer in a ratio of 1:1. As the
smectite-like layers in REC are readily intercalated, REC possesses
* Corresponding author. Tel.: +86 22 27406144; fax: +86 22 27403475.
E-mail address: maxiaofei@tju.edu.cn (X. Ma).
good adsorption capacity for cationic metals and dyes [23–26] by
exchanging interlayer Na+ with either organic or inorganic cations.
Previous work proved that modified REC is a good adsorbent for
MB (methylene blue) and Pb2+, with maximum adsorption values
of 277 mg/g and 180.8 mg/g, respectively [27].
However, REC exhibited weak adsorption for anionic contami-
nants such as nitrate, phosphate, sulfate and dichromate. Therefore,
it is of great necessity to modify REC in order to meet the demand for
the removal of nitrate and phosphate contaminants in water.
Recently, many studies focused on tackling such similar problem,
and quaternary ammonium salt is a very commonly used and
advisable modifier [28]. By ion exchange between organic cations
and Na+ in the interlayer of REC, the adsorption capacity of REC for
anionic contaminants can be greatly enhanced. In addition, another
disadvantage of REC adsorbent is its difficult withdrawal from water
after the adsorption process. By introducing magnetic Fe3O4 into
REC, REC composites can be easily separated from water with the
magnetic field, thus avoiding secondary pollution [29]. In this work,
magnetic Fe3O4 was introduced into the interlayer of REC for the
facile separation from water. At the same time, Fe3O4 particles could
improve the interlayer spaces of REC, which would be beneficial for
intercalating quaternary ammonium salt into the enlarged inter-
layer of REC. The REC/Fe3O4 composite was then treated with CTAB
to obtain REC/Fe3O4-CTAB composite. The modifications were
expected to improve the adsorption capacities for nitrate (NO3)
and phosphate (PO43). In this process, magnetic REC actually acted

http://dx.doi.org/10.1016/j.jiec.2016.07.017
1226-086X/ß 2016 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.
166 F. Wang et al. / Journal of Industrial and Engineering Chemistry 41 (2016) 165–174
as the support of CTAB, while CTAB on magnetic REC was the
adsorbent for nitrate and phosphate.
Experimental
Materials
Sodium rectorite was purchased by Hubei Zhongxiang Rectorite
Mine (Wuhan, China). All other reagents were of analytical grade
and commercially available.
Preparation of REC/Fe3O4 composite
REC/Fe3O4 composites were fabricated in a modified method of
Ref. [29]. 1 g REC, 1.165 g FeCl36H2O and 0.6 g FeSO47H2O were
added into 20 mL distilled water. 20 mL NH3H2O solution (8 M)
was added dropwise to produce magnetic iron oxide under N2
atmosphere. The pH of the mixture was approximately 11–12. The
suspension was held at 70 8C for 4 h and then washed with distilled
water several times. The obtained REC/Fe3O4 composite was dried
at 50 8C for 4 h.
Preparation of REC-CTAB composites
REC-CTAB composites were prepared in a modified method of
Ref. [30]. 4 g REC was dispersed in 200 mL 0.014 mol/L CTAB
solution using ultrasonication for 10 min. The pH of the mixture
was adjusted to 2, 7 and 13 with hydrochloric and sodium
hydroxide solution. Then the mixture was shaken on an immersion
oscillator registration at 100 rpm for 12 h at 60 8C. After washed for
6–8 times, the sample was dried at 50 8C for 4 h. The obtained REC-
CTAB composites were labeled as CTAB-REC-2, CTAB-REC-7 and
CTAB-REC-13, respectively.
Assembly of REC-CTAB/Fe3O4 composite
1 g as-obtained CTAB-REC-13, 1.165 g FeCl36H2O and 0.6 g
FeSO47H2O were added into 20 mL distilled water, and 20 mL
NH3H2O solution (8 M) was added dropwise to produce magnetic
iron oxide under N2 atmosphere. The pH of the mixture was
approximately 11–12. The suspension was held at 70 8C for 4 h and
then washed with distilled water several times. The obtained REC-
CTAB/Fe3O4 composite was dried at 50 8C for 4 h.
Assembly of REC/Fe3O4-CTAB composite
1 g as-obtained REC/Fe3O4 composite was dispersed in 50 mL
0.014 mol/L CTAB solution using ultrasonication for 10 min. The
pH of the mixture was adjusted to 13 with sodium hydroxide
solution. Then the mixture was shaken on an immersion oscillator
registration at 100 rpm for 12 h at 60 8C. After washed for 6–8
times, the obtained REC/Fe3O4-CTAB composite was dried at 50 8C
for 4 h.
X-ray diffraction (XRD)
X-ray diffraction patterns for REC, REC-CTAB, REC/Fe3O4, REC-
CTAB/Fe3O4 and REC/Fe3O4-CTAB were recorded in reflection mode
in the angular range of 1–408 (2u), at room temperature by a D/
MAX-2500 operated at a CuKa wavelength of 1.542 Å.
Fourier transform infrared spectroscopy (FTIR) spectra
FTIR analysis of REC, REC-CTAB, REC/Fe3O4, REC-CTAB/Fe3O4
and REC/Fe3O4-CTAB were performed on a BIO-RAD FTS3000 IR
spectra scanner. The sample powders were dispersed in KBr and
pressed into transparent sheets for testing.
Thermogravimetric analysis (TGA)
Thermal properties of REC, REC-CTAB, REC/Fe3O4, REC-CTAB/
Fe3O4 and REC/Fe3O4-CTAB were measured using a ZTY-ZP type
thermal analyzer. Sample weights ranged from 10 to 15 mg.
Samples were heated from ambient temperature to 800 8C at a
heating rate of 10 8C/min in a nitrogen atmosphere.
Vibrating sample magnetometer (VSM)
The magnetic properties of REC/Fe3O4, REC-CTAB/Fe3O4, and
REC/Fe3O4-CTAB were tested on a vibrating sample magnetometer
(LDJ 9600-1, LDJ Electronics Inc., USA).
Scanning electron microscope (SEM)
Powders of REC, REC-CTAB, REC/Fe3O4, REC-CTAB/Fe3O4 and
REC/Fe3O4-CTAB were immobilized in a sample holder, coated
with gold, and viewed with an S-4800 scanning electron
microscope.
Transmission electron microscope (TEM)
REC, REC-CTAB, REC/Fe3O4, REC-CTAB/Fe3O4, REC/Fe3O4-CTAB
powders were dispersed in alcohol. Sample suspensions were
dropped into a copper grid, dried in air, and tested with a JEM-
2100F transmission electron microscope.
Energy dispersive spectroscopic (EDS) mapping
The composition of REC-CTAB/Fe3O4 and REC/Fe3O4-CTAB was
measured with energy dispersive spectroscopic mapping with a
JEM-2100F transmission electron microscope.
Specific surface area
The specific surface area of REC, REC-CTAB/Fe3O4 and REC/
Fe3O4-CTAB were measured with an Autosorb-1 specific surface
area analyzer (Quantachrome Instruments, USA) using the BET
method on the base of N2 adsorption data.
The point of zero charge (pHZPC)
The point of zero charge (pHZPC) of adsorbent REC/Fe3O4-CTAB
was determined. 10 mg REC/Fe3O4-CTAB composite was added to
10 mL of 0.1 mol/L KNO3 solution at different initial pH from 3 to
10. The initial pH of solutions were adjusted with NaOH or HCl
solution, and measured by pH meter (PHSJ-4A, China). Afterwards,
the mixtures were shaken on a rotary shaker at 100 rpm for 12 h at
30 8C and the final pH values of solutions were measured at
equilibrium [31].
Adsorption of nitrate (NO3)
Adsorption experiments were conducted using glass bottles
containing 10 mg adsorbents and 10 mL nitrate solution (NO3
100 mg/L) at pH 5. During the adsorption process, the glass
bottles were placed on a slow-moving platform shaker and
nitrate solutions were taken at intervals to analyze the effect of
contact time on equilibrium adsorption capacity. The NO3
concentrations were tested using UV–vis spectrometry at
204 nm to determine the adsorbed amounts of NO3 at 30 8C
[32].
 [(Fig._1)TD$IG]
F. Wang et al. / Journal of Industrial and Engineering Chemistry 41 (2016) 165–174 167

In adsorption isotherm testing, the concentration of the (a)

adsorbent was 1 g/L and the NO3 concentrations varied from
80 to 500 mg/L. These experiments were carried out at 30 8C and all
suspensions were shaken on a rotary shaker at 100 rpm for 2 h to
reach the adsorption equilibrium.

Adsorption of phosphate (PO43) CTAB-REC-13

Intensity
Adsorption experiments were conducted using glass bottles CTAB-REC-7
containing 1 g/L adsorbents and 10 mL phosphate solution (PO43
100 mg/L) at pH 5 (mainly in the form of H2PO42). Solutions of
ammonium molybdate-sulfuric acid and stannous chloride were CTAB-REC-2
added to show blue color. During the adsorption process, the glass
R(001) REC R(002)
bottles were placed on a slow-moving platform shaker and
phosphate solution were taken at intervals to study the effect of
contact time on equilibrium adsorption capacity. The PO43
2 4 6 8 10
concentrations were tested using UV–vis spectrometry at
650 nm to determine the adsorbed amounts of PO43 at 30 8C. 2 theta (degree)
In adsorption isotherm testing, the concentration of the
adsorbent was 1 g/L and the PO43 concentrations varied from (b)
80 to 500 mg/L. These experiments were carried out at 30 8C and all
suspensions were shaken on a rotary shaker at 100 rpm for 2 h to
reach the adsorption equilibrium.
REC/Fe 3O4-CTAB
The regeneration of REC/Fe3O4-CTAB

For the readsorption study of NO3 and PO43–, 20 mg REC/ Intensity REC-CTAB/Fe 3O4
Fe3O4-CTAB was added to 20 mL NO3 (100 mg/L) and PO43
(100 mg/L) solution respectively, and placed on a rotary shaker at
R(001)
100 rpm for 2 h. REC/Fe3O4-CTAB was taken out from the solution,
and the concentrations of NO3 and PO43 in the solution were REC/Fe 3O4
determined with a UV–vis spectrometer. REC/Fe3O4-CTAB was R(002)
dried and immersed in 20 mL NaOH (1 mol/L) solution for 12 h to
regenerate. The obtained REC/Fe3O4-CTAB was washed with REC
distilled water for 4–6 times and readsorb at pH 5 (adjusted with
1 mol/L HCl solution). 2 4 6 8 10
2 theta (degree)
Results and discussion (c)

Characterization of REC/Fe3O4-CTAB composites

REC/Fe 3O4-CTAB
As shown in Fig. 1(a), raw REC displayed strong R(0 0 1) and
R(0 0 2) diffraction peaks at 2u = 4.108 and 8.118. According to the
REC-CTAB/Fe 3O4
Bragg diffraction equation, 2dsinu = l, the distance between the
REC layers (d0 0 1 and d0 0 2) can be approximately 2.16 and
intensity

1.09 nm. After being modified by CTAB, the R(0 0 1) patterns F(311)
shifted to 3.128, 2.848 and 2.788 for CTAB-REC-2, CTAB-REC-7 and
REC/Fe 3O4 F(220)
CTAB-REC-13, respectively, and their interlayer distance expanded
to 2.83, 3.11 and 3.18 nm. Moreover, R(0 0 2) patterns shifted to
the lower 2 theta values. These indicated that CTAB successfully
intercalated the layers of REC, and cationic exchange reaction REC
occurred between REC and CTAB. As CTAB-REC-13 expanded most
obviously, REC modifications by CTAB were most effective at pH
13. On the one hand, REC showed better dispersion in basic 10 20 30 40
solution, which facilitated better interaction between REC layers 2 theta (degree)
and CTAB. On the other hand, at pH 13 the surface of REC would
become negatively charged, which would in turn favors the Fig. 1. X-ray diffraction peaks of REC-CTAB prepared at the different pH values (a),
and REC/Fe3O4, REC-CTAB/Fe3O4, REC/Fe3O4-CTAB (from 1 to 108) (b) and (from 1 to
positively charged quaternary ammonium cations and enhance
408) (c).
the cationic exchange between REC and CTAB [33]. Therefore, the
synthesis of REC-CTAB/Fe3O4 and REC/Fe3O4-CTAB involved
adjusting pH value to 13 using sodium hydroxide solution. However, the intensity of R(0 0 2) reflection was obviously
As shown in Fig. 1(b), after being modified by iron oxide, the weakened. This suggested that the introduction of Fe3O4 partially
R(0 0 1) shifted to the lower values of 2 theta for REC/Fe3O4, REC- damaged the ordered structure of REC, and that the layers of REC
CTAB/Fe3O4 and REC/Fe3O4-CTAB. It indicated that the cationic were intercalated or exfoliated with different degrees. And CTAB
exchange occurred between Fe3+/Fe2+ and Na+ in the layers of REC, could further intercalate or exfoliate the REC layer. However, Fe3O4
and the formation of Fe3O4 occurred at the surface of REC. hardly improved interlayer distance if REC was firstly intercalated
168 [(Fig._3)TD$IG]
F. Wang et al. / Journal of Industrial and Engineering Chemistry 41 (2016) 165–174

by CTAB. Therefore, the order of the introduction of Fe3O4 and CTAB (a)
had great effect on the intercalation of REC.

Mass loss (%)

As observed in Fig. 1(c), the characteristic diffraction peaks of 100
REC-Fe3O4 at 30.2 and 35.7 are related to [2 2 0] and [3 1 1] crystal
planes of face-centered cubic Fe3O4. These two peaks are also seen
in REC-CTAB/Fe3O4 and REC/Fe3O4-CTAB, confirming that iron REC
oxide has been grafted and attached on REC layers [34]. CTAB-REC-2
The FTIR spectra of REC, REC-CTAB, REC/Fe3O4, REC-CTAB/Fe3O4 80 CTAB-REC-7
and REC/Fe3O4-CTAB were shown in Fig. 2. In FTIR spectra of REC CTAB-REC-13
[35], the peak at 3644 cm1 was related to the hydroxyl stretching
bonding of Si–O–H. The peak at 1021 cm1 was associated to the 0

DTG
in-plane Si–O–Si stretching vibration, and the peaks at 818 and
703 cm1 were respectively attributed to Si–O–Al bending and Al–
O– out of plane. These peaks appeared in the FTIR spectra of REC- 421
262
CTAB, REC/Fe3O4, REC-CTAB/Fe3O4 and REC/Fe3O4-CTAB. After REC
was modified by CTAB, there appeared two new peaks at 2850 and -2
200 400 600 800
2926 cm1 for REC-CTAB. These new peaks were attributed to the
symmetric and asymmetric CH2– stretching vibration, which could
also be seen in REC-CTAB/Fe3O4 and REC/Fe3O4-CTAB [36]. (b)
However, CH2– stretching vibration was much weaker in REC-

Mass loss (%)

CTAB/Fe3O4. This was related to the overlapping by Fe3O4, 100
weakening the effect of CTAB.
According to Fig. 3, the thermogravimetric (TG) curves of REC
exhibited no weight loss under 500 8C. For REC-CTAB, the weight-
REC
loss curve decreased continuously after 200 8C in Fig. 3(a). The
80 REC-CTAB/Fe3O4
weight loss at about 262 8C was attributed to thermal decomposi-
tion of the quaternary ammonium cations on the REC layers [30]. REC/Fe3O4-CTAB
At 421 8C, the weight loss could be related to the pyrolysis of the
long alkyl chains of CTAB [37], and the interaction between CTAB 0
and REC layers improved the decomposed temperature of CTAB.
DTG

265
Obviously, at pH 13 more CTAB were loaded in REC layers, in 411
accordance with the highest weigh loss peak at 262 8C in Fig. 3(a). 335
The thermogravimetric (TG) curves of REC-CTAB/Fe3O4 and REC/
Fe3O4-CTAB were shown in Fig. 3(b). As for REC-CTAB/Fe3O4, the -2
loss of free water and the decomposition of iron hydroxide 200 400 600 800

impurity could result in the weight loss below 200 8C and at 335 8C,
respectively. In views of weight loss at 411 8C, a higher weight loss
Fig. 3. TG curves of REC-CTAB prepared at the different pH values (a), and REC-
in REC/Fe3O4-CTAB was observed because more CTAB was attached CTAB/Fe3O4, REC/Fe3O4-CTAB (b).
on REC in REC/Fe3O4-CTAB than REC-CTAB/Fe3O4.
As REC can disperse well in water, it is necessary to separate raw
REC or modified REC in determining residual nitrate or phosphate 300 K. REC/Fe3O4 and REC-CTAB/Fe3O4 composites possessed
or reclaiming REC. Fig. 4 displays the magnetic hysteresis curves of magnetic properties with saturation magnetization (18.28 and
REC/Fe3O4, REC-CTAB/Fe3O4 and REC/Fe3O4-CTAB measured at 17.55 emu/g, respectively) and exhibited superparamagnetic
[(Fig._2)TD$IG] behavior with small coercivity in the applied magnetic fields
[(Fig._4)TD$IG]
REC/Fe3O4-CTAB
2852 20
2922 1401 696
3641 3136 REC-Fe3O4
REC-CTAB/Fe3O4 15
1021 REC-CTAB/Fe3O4
3641 1401 685
3130 10 REC/Fe3O4-CTAB
1021
magnetization (emu/g)

REC/Fe3O4 684
3642
Transmittance

1403 5
3126
1022
3645 2926 2850 REC-CTAB 1479 702 0
REC 1058
3046 -5
1440 703
-10
3644
-15

1021 -20

4000 3500 3000 2500 2000 1500 1000 500 -10000 -5000 0 5000 10000
Wavenumber (cm-1) Field (Oe)

Fig. 2. FTIR spectra of REC, REC-CTAB, REC/Fe3O4, REC-CTAB/Fe3O4 and REC/Fe3O4- Fig. 4. Hysteresis loops of REC/Fe3O4, REC-CTAB/Fe3O4, REC/Fe3O4-CTAB measure at
CTAB. 300 K.
 F. Wang et al. / Journal of Industrial and Engineering Chemistry 41 (2016) 165–174
[38]. REC/Fe3O4-CTAB exhibited a smaller saturation magnetiza-
tion value (8.49 emu/g) than REC/Fe3O4. The magnetism of Fe3O4
was reduced in the process of CTAB introduction. Nevertheless, the
saturation magnetization values of them were enough to separate
them from aqueous solution with a strong magnet.
SEM images were shown in Fig. 5. The aggregated layers of REC
(Fig. 5(a)) were identified. After being modified by CTAB, the layers
of REC were intercalated, and became obvious, as marked with red
circles in Fig. 5(b and c). In Fig. 5 (d, e and f), Fe3O4 particles were
introduced and grafted onto the surface of the layers in REC [39].
As shown in Fig. 6, REC possessed platelets with a thickness of
30–100 nm and the layers were transparent. After modification
with quaternary ammonium salt, the CTAB molecules were easily
visualized on the surface of REC layers via contrast differences
between REC and REC-CTAB as some scattered dots could be seen
in Fig. 6(b). With the introduction of iron oxide, Fe3O4 particles
could be easily observed with a diameter of 40–90 nm in Fig. 6(c
and d). After modification, some REC stacks were still consisted of
[(Fig._5)TD$IG]
Fig. 5. SEM images of REC (a), REC-CTAB (b and c) , REC/Fe3O4 (d), REC-CTAB/Fe3O4 (e) and REC/Fe3O4-CTAB (f). (For interpretation of the references to color in the text, the
reader is referred to the web version of this article.)
169
several layers, while some REC layers were exfoliated, and the REC
was thinner [40].
In order to determine the composition of modified REC, energy
dispersive spectroscopic (EDS) mapping was carried out [41]. Fig. 7
confirmed that the quaternary ammonium groups and Fe3O4
particles were successfully introduced onto the layers of REC.
Besides, CTAB and Fe3O4 particles could be homogeneously
dispersed on REC surface.
The BET surface area of REC was 9.94 m2/g [42]. Nevertheless,
the BET surface areas of REC/Fe3O4 and REC/Fe3O4-CTAB were
406.05 and 434.42 m2/g, respectively, much higher than that of
REC. The increased surface area of REC/Fe3O4 composite was
attributed to the intercalation of REC layers and magnetic iron
oxide particles. When CTAB was introduced onto the REC interlayer
afterward, the basal spacing of REC was expanded further,
implying that CTAB was definitely grafted onto the layers of REC.
The point of zero charge (pHZPC) of REC/Fe3O4-CTAB composite
was shown in Fig. 8. It could be noted that pHZPC, namely the pH at
[(Fig._6)TD$IG]
170 F. Wang et al. / Journal of Industrial and Engineering Chemistry 41 (2016) 165–174

Fig. 6. TEM micrographs for REC (a), REC-CTAB (b), REC-CTAB/Fe3O4 (c) and REC/Fe3O4-CTAB (d).

[(Fig._7)TD$IG]

Fig. 7. Energy dispersive X-ray analysis and dot mapping on REC-CTAB/Fe3O4 (a) and REC/Fe3O4-CTAB (b).
[(Fig._8)TD$IG] [(Fig._9)TD$IG]
F. Wang et al. / Journal of Industrial and Engineering Chemistry 41 (2016) 165–174 171

11
(a)
pH (initial) 80
10
pH (Final) REC/Fe 3O4-CTAB
9
60 REC-CTAB
8
pH (Final)

qt (mg/g)
7
40
6 REC-CTAB/Fe 3O4

5 20 REC

3 0

2
0 20 40 60 80 100 120
2 3 4 5 6 7 8 9 10 11
pH (Initial) t (min)

Fig. 8. Determination of the pH of point of zero charge (pHZPC). 60

(b) REC/Fe 3O4-CTAB

which the net surface charge on REC/Fe3O4-CTAB composite was 50 REC-CTAB

zero (DpH 0) was associated with a pH value around 5.4. It means
that at pH values less that 5.4, the REC/Fe3O4-CTAB surface had a
40
net positive charge, while at pH above 5.4, the surface had a net REC-CTAB/Fe 3O4
negative charge. Therefore, the acidic pH could enhance the
adsorption of NO3 and PO43 onto REC/ Fe3O4-CTAB surface [31]. 30
qt (mg/g)

However, if the solution was excessively acidic, molecular form of REC

H3PO4 would prevail, which would damage the adsorption for 20
phosphate.
10
Adsorption Kinetic of NO3 and PO43

Fig. 9 reveals the effect of contact time on adsorption of NO3 0

and PO43 by REC, REC-CTAB, REC-CTAB/Fe3O4 and REC/Fe3O4-
CTAB. For about 2 h, the adsorptions could reach the equilibrium. 0 20 40 60 80 100 120
In Fig. 9, CTAB possessed a higher adsorption capacity for NO3 and t (min)
PO43 than raw REC because the grafted positively-charged CTAB
Fig. 9. (a) Effect of contact time on adsorption of NO3 by REC, REC-CTAB, REC-CTAB/
generated electrostatic attraction towards anions. It was obvious
Fe3O4 and REC/Fe3O4-CTAB at 30 8C. Initial concentration: NO3 100 mg/L;
that REC/Fe3O4-CTAB had absolute predominance whether in the adsorbents 1 g/L. (b) Effect of contact time on adsorption of PO43 by REC, REC-
nitrate adsorption or phosphate adsorption. However, the adsorp- CTAB, REC-CTAB/Fe3O4 and REC/Fe3O4-CTAB at 30 8C. Initial concentration: PO43
tion capacity for NO3 and PO43 of REC-CTAB/Fe3O4 was far less 100 mg/L; adsorbents 1 g/L.
than that of REC/Fe3O4-CTAB, even though they went thorough
same procedures in different order. The reason could be attributed displayed larger adsorption ability for NO3 and PO43 than raw
to the chemical site occupation of Fe3O4 on the adsorption sites of REC.
REC, damaging the adsorption performance of quaternary ammo- The adsorption process can be measured using pseudo-second-
nium cations on the layers of REC. Besides, as Fe3O4 particles order kinetic model. According to this model, the amount of NO3
contributed little to the adsorption for anions, the introduction of and PO43 adsorbed at any time, t (min), is defined as [43]:
Fe3O4 increased the weight of adsorbents and decreased the t 1 t
adsorption capacity of 1 g adsorbents when iron oxide was grafted ¼ þ (1)
qt kqe af;2 qe
after CTAB. However, when Fe3O4 was firstly introduced before
CTAB, it would enhance the interlayer spaces of REC, thus allowing where k (mg/g min1) is the second order rate constant, and qt
more CTAB molecules enter REC layers and facilitating higher (mg/g) and qe (mg/g) represent the amount of NO3 and PO43
adsorption capacity for anions [36]. That could be verified in adsorbed at any time (min) and equilibrium, respectively.
Fig. 3(b) that REC/ Fe3O4-CTAB experienced a higher weight loss Kinetic constants were calculated from the experimental data
peak at 411 8C than REC-CTAB/Fe3O4. But the overall modified REC in Fig. 9, and listed in Tables 1 and 2. The adsorption process of

Table 1
Kinetic constants for NO3 adsorption at 30 8C.

Pseudo-second-order model Intra-particle diffusion model

qe (mg/g) k (g/mg min) R kid (mg/g min1/2) C R

REC 19.2 0.015 0.9994 0.65 12.36 0.9354

REC-CTAB 58.2 0.0038 0.9998 2.52 32.13 0.9195
REC-CTAB/Fe3O4 34.3 0.0019 0.9984 2.31 6.24 0.9452
REC/Fe3O4-CTAB 87.0 0.0012 0.9944 3.02 39.85 0.9444
172 F. Wang et al. / Journal of Industrial and Engineering Chemistry 41 (2016) 165–174

Table 2
Kinetic constants for PO43 adsorption at 30 8C.

Pseudo-second-order model Intra-particle diffusion model

qe (mg/g) k (g/mg min) R kid (mg/g min1/2) C R

REC 25.3 0.0047 0.9996 1.46 9.39 0.9363

REC-CTAB 51 0.0044 0.9997 2.08 29.17 0.9250
REC-CTAB/Fe3O4 38.1 0.0043 0.9993 1.94 17.71 0.9219
REC/Fe3O4-CTAB 59.3 0.0019 0.9999 3.63 20.11 0.9233

[(Fig._10)TD$IG]
NO3 and PO43 matched the pseudo second-order model well 200
(a)
with a high correlation coefficient (R > 0.992), and qe values of
180
REC-CTAB, REC-CTAB/Fe3O4 and REC/Fe3O4-CTAB were 58.2, 34.3, REC/Fe 3O4-CTAB
87.0 mg/g for NO3 and 51.0, 38.1, 59.3 mg/g for PO43, 160
respectively.
140
The intra-particle diffusion model was also widely applied to
measure the diffusion mechanism of adsorption process. It is 120 REC-CTAB

qe (mg/g)
defined as [25]:
100
qt ¼ K id t 1=2 þ C (2)
80
where Kid (mg/g min1/2) is the intra-particle diffusion rate constant REC-CTAB/Fe 3O4
60
and C is the intercept. REC
As shown in Tables 1 and 2, the pseudo-second-order kinetic 40
model fitted the adsorption process of NO3 and PO43 better than
the intra-particle diffusion model, as the pseudo-second-order 20
kinetic model exhibited higher R values. 0 100 200 300 400 500
Ce (mg/L)
Adsorption isotherms of NO3 and PO43
(b)
160
Adsorption isotherms of NO3 and PO43 by REC, REC-CTAB,
REC/Fe 3O4-CTAB
REC-CTAB/Fe3O4 and REC/Fe3O4-CTAB were shown in Fig. 10, and
investigated with the Langmuir model, expressed as [43]: 140

ce 1 ce 120
¼ þ (3) REC-CTAB
qe bQ max Q max
qe (mg/g)

100
where ce (mg/L) is the equilibrium concentration in the solution; qe
(mg/g) is the equilibrium adsorption capacity. Qmax (mg/g) is the 80
maximum adsorption capacity; b (L/mg) is the Langmuir adsorp- REC-CTAB/Fe 3O4
tion equilibrium constant.
60
The adsorption isotherm process of NO3 and PO43 by REC, REC
REC-CTAB, REC-CTAB/Fe3O4 and REC/Fe3O4-CTAB (Fig. 10) fitted
40
the Langmuir isotherm model well, and the maximum monolayer
adsorption capacities (Qmax) for REC-CTAB, REC-CTAB/Fe3O4 and
20
REC/Fe3O4-CTAB were 134.4, 126.7, 182.1 mg/g for NO3 and 0 100 200 300 400 500
152.9, 117.0, 174.5 mg/g for PO43, respectively, as listed in Tables
Ce (mg/L)
3 and 4. Above all, it was easily observed that REC/Fe3O4-CTAB was
an excellent adsorbent with predominant maximum adsorption Fig. 10. (a) Adsorption isotherms for NO3 adsorption by REC, REC-CTAB, REC-
capacity for both NO3 (182.1 mg/g) and PO43 (174.5 mg/g). The CTAB/Fe3O4 and REC/Fe3O4-CTAB at 30 8C. Adsorbents 1 g/L. (a) Adsorption
Qmax values of REC/Fe3O4-CTAB for nitrate (NO3) were competi- isotherms for PO43 adsorption by REC, REC-CTAB, REC-CTAB/Fe3O4 and REC/
tive with other adsorbents (Table 5) such as polyacrylonitrile- Fe3O4-CTAB at 30 8C. Adsorbents 1 g/L.

alumina nanoparticle mixed matrix membrane (15 mg/g) [43],

modified pine sawdust (32.8 mg/g) [44], quaternized chitosan-
melamine-glutaraldehyde resin (97.5 mg/g) [45], Zr(IV) loaded

Table 3
Isothermal constants for NO3 adsorption at 30 8C.

Langmuir model Freundlich model

Qmax (mg/g) b (L/mg) R 1/n KF (mg11/n L1/n g1) R

REC 82.4 0.0034 0.9956 0.62 1.21 0.9981

REC-CTAB 134.4 0.014 0.9956 0.37 13.4 0.9927
REC-CTAB/Fe3O4 126.7 0.0046 0.9955 0.60 2.35 0.9860
REC/Fe3O4-CTAB 182.1 0.0422 0.9999 0.29 34.8 0.9802
 F. Wang et al. / Journal of Industrial and Engineering Chemistry 41 (2016) 165–174 173

Table 4
Isothermal constants for PO43 adsorption at 30 8C.

Langmuir model Freundlich model

Qmax (mg/g) b (L/mg) R 1/n KF (mg11/n L1/n g1) R

REC 61.6 0.0117 0.9992 0.36 5.99 0.9822

REC-CTAB 152.9 0.0102 0.9936 0.49 7.18 0.9694
REC-CTAB/Fe3O4 117 0.0077 0.9928 0.52 4.30 0.9714
REC/Fe3O4-CTAB 174.5 0.0226 0.9992 0.37 19.5 0.9705
[(Fig._1)TD$IG]
Table 5 NO3- adsorption
The comparison of maximum adsorption capacity for NO3 and PO43. 88.3
84.5 PO43-adsorption
Adsorbents Adsorption capacities 79.3
(mg g1) 80
74.2 72.6
S
NO3 adsorption
Polyacrylonitrile-alumina nanoparticle 15 mg/g [43]
65.0
mixed matrix membrane 60

efficiency (%)
Modified pine sawdust 32.8 mg/g [44]
Quaternized chitosan-melamine- glutaraldehyde 97.5 mg/g [45]
resin
Zr(IV) loaded cross-linked chitosan beads 128.4 mg/g [46] 40
REC/Fe3O4-CTAB composite 182.1 mg/g [this work]
PO43S adsorption
Mg–Al hydrotalcite-loaded kaolin clay 11.9 mg/g [47]
Fe–Ti bimetal oxide 35.4 mg/g [48] 20
Modified chitosan beads 60.6 mg/g [49]
Quaternized chitosan-melamine-glutaraldehyde 112.5 mg/g [45]
resin
Zr(IV) loaded cross-linked chitosan beads 149.4 mg/g [46] 0
REC/Fe3O4-CTAB composite 174.5 mg/g [this work] 1 2 3
cycle time
cross-linked chitosan beads (128.4 mg/g) [46]. Also, the adsorption Fig. 11. Three consecutive adsorption cycles of REC/Fe3O4-CTAB for NO3 and
capacities of REC/Fe3O4-CTAB for phosphate (PO43) were higher PO43.
than other adsorbents, including Mg–Al hydrotalcite-loaded kaolin
clay (11.9 mg/g) [47], Fe–Ti bimetal oxide (35.4 mg/g) [48], Langmuir model exhibited better fitness than the Freundlich
modified chitosan beads (60.6 mg/g) [49], quaternized chitosan- model. The 1/n values were less than 1, so the isotherm adsorption
melamine-glutaraldehyde resin (112.5 mg/g) [45], and Zr(IV) was heterogeneous [49].
loaded cross-linked chitosan beads (149.4 mg/g) [46].
The Freundlich model was also widely used to describe The regeneration of REC/Fe3O4-CTAB
adsorption isotherm, and expressed as [50]:
The adsorption cycles of REC/Fe3O4-CTAB were shown in Fig. 11.
1 The adsorption efficiency (%) meant the ratio of the weight of
log qe ¼ logK F þ logC e (4)
n adsorbed NO3 and PO43 at any time and that at the first time. At a
where 1/n is the Freundlich adsorption intensity parameter and KF higher pH, the surface of adsorbents could carry more negative
(mg11/n L1/n g1) is the Freundlich constant. As shown in Tables 3 charges and therefore would more significantly repulse the
and 4, as for the adsorption process of NO3 and PO43, the negatively charged ions in solution, such as NO3 and PO43.
[(Fig._12)TD$IG]

Fig. 12. Mechanism of NO3 and PO43 removal.

174 F. Wang et al. / Journal of Industrial and Engineering Chemistry 41 (2016) 165–174

The adsorption efficiency of NO3 and PO43 reduced with the Provincial Education Department (L2013505), and the National
increasing of adsorption cycles. The adsorption efficiency of NO3 Nature Science Foundation of China (51462009).
and PO43 on REC/Fe3O4-CTAB decreased to 65.0 and 72.6% at the
third time. After several circles, REC/Fe3O4-CTAB still had good References
adsorption capacity even though the adsorption efficiency
decreased much. [1] P. Loganathan, S. Vigneswaran, J. Kandasamy, J. Environ. Manag. 131 (2013) 363.
[2] H. Song, Y. Zhou, A. Li, S. Mueller, Desalination 296 (2012) 53.
[3] O. Primo, M.J. Rivero, A.M. Urtiaga, I. Ortiz, J. Hazard. Mater. 164 (2009) 389.
Mechanism of nitrate and phosphate removal [4] A. Pintar, J. Batista, J. Levec, Chem. Eng. Sci. 56 (2001) 1551.
[5] J.H. Ahn, K.H. Choo, H.S. Park, J. Membr. Sci. 310 (2008) 296.
The mechanism of nitrate and phosphate removal by REC/ [6] H.H. Hwang, D.G. Kim, H.S. Shin, J. Hazard. Mater. 185 (2011) 1513.
[7] M.Z. Kassaee, E. Motamedi, A. Mikhak, R. Rahnemaie, Chem. Eng. J. 166 (2011) 490.
Fe3O4-CTAB composite was shown in Fig. 12. The fabrication of [8] L. Jiang, L. Lu, W. Zhang, Q. Du, B. Pan, L. Yang, Q. Zhang, Water Res. 45 (2011) 2191.
REC/Fe3O4-CTAB composite involved two procedures: the interca- [9] Y.H. Liou, S.L. Lo, W.H. Kuan, C.J. Lin, S.C. Weng, Water Res. 40 (2006) 2485.
lation of REC by magnetic iron oxide particles and cationic [10] J.S. Shengnan Yi, H. He, C. Long, A. Li, Chem. Eng. J. 230 (2013) 166.
[11] M. Kumar, S. Chakraborty, J. Hazard. Mater. 135 (2006) 112.
exchange between REC and cationic surfactant (i.e. CTAB). The [12] S.M. Hosseini, B. Ataie-Ashtiani, M. Kholghi, Desalination 276 (2011) 214.
introduction of Fe3O4 particles expanded the interlayer spacing of [13] Y.A. Tielong Li, Z. Jin, M. Dong, H. Xia, X. Wang, Bioresour. Technol. 101 (2010) 9825.
REC and therefore permitted more CATB molecules to enter REC [14] P.T. Chiueh, Y.H. Lee, C.Y. Su, S.L. Lo, J. Hazard. Mater. 192 (2011) 837.
[15] X. Xing, B.Y. Gao, Q.Q. Zhong, Q.Y. Yue, Q. Li, J. Hazard. Mater. 186 (2011) 206.
layers and contributed to a higher adsorption capacity for nitrate [16] A. Bhatnagar, M. Sillanpaa, Chem. Eng. J. 168 (2011) 493.
and phosphate anions. The grafted positively-charged CTAB [17] A. Maleki, B. Hayati, M. Naghizadeh, S.W. Joo, J. Ind. Eng. Chem. 28 (2015) 211.
molecules in REC layers could cause electrostatic attraction [18] K. Mizuta, T. Matsumoto, Y. Hatate, K. Nishihara, T. Nakanishi, Bioresour. Technol.
95 (2004) 255.
towards anions, such as NO3 and PO43 [50]. [19] H.L. Jiang, P.H. Chen, S.L. Luo, Appl. Clay Sci. 284 (2013) 942.
[20] S.L. Tian, P.X. Jiang, P. Ning, Chem. Eng. J. 151 (2009) 141.
[21] L. Borgnino, M.J. Arena, C.P. Pauli, Colloid Surf. A 341 (2009) 46.
Conclusion [22] J. Matusik, T. Bajda, J. Colloid Interface Sci. 398 (2013) 74.
[23] L.L. Yang, X.Y. Ma, N.N. Guo, Carbohydr. Polym. 90 (2012) 853.
[24] L.X. Zeng, Y.F. Chen, Q.Y. Zhang, Carbohydr. Polym. 130 (2015) 333.
The raw REC was successfully modified by introducing Fe3O4 [25] Y.J. Lu, P.R. Chang, P.W. Zheng, Appl. Clay Sci. 107 (2015) 21.
particles and CTAB cations. During the modifications, iron oxide [26] G.K. Zhang, Y. Liu, X. Yang, Appl. Clay Sci. 29 (2005) 15.
first formed in the layers of REC, and then CTAB easily intercalated [27] F. Wang, P.R. Chang, P.W. Zheng, X.F. Ma, Appl. Surf. Sci. 349 (2015) 251.
[28] H.L. Hong, W.T. Jiang, X.L. Zhang, Appl. Clay Sci. 42 (2008) 292.
the expanded REC interlayer. CTAB on magnetic REC could remove [29] Y.F. Xie, D.Y. Qian, D.L. Wu, X.F. Ma, Chem. Eng. J. 168 (2011) 959.
nitrate and phosphate. Besides, it is convenient to reclaim [30] J.H. Wang, X. Zhang, B. Zhang, Desalination 259 (2010) 22.
adsorbents from the agricultural wastewater with the existence [31] S.S. Mehdi, A. Khataee, A. Hassani, S. Karaca, C.R. Chim. 18 (2015) 204.
[32] Y.J. Lu, P.R. Lu, P.W. Zheng, X.F. Ma, Chem. Eng. J. 255 (2014) 49.
of magnetic Fe3O4 particle on the layers of REC. The adsorption
[33] W.B. Wang, A.Q. Wang, Carbohydr. Polym. 77 (2009) 891.
process of nitrate and phosphate matched the pseudo second- [34] L.X. Zheng, M.J. Xie, Q.Y. Zhang, Carbohydr. Polym. 123 (2015) 89.
order model and Langmuir isotherm model well. The maximum [35] G.C. Han, Y. Han, X.Y. Wang, Appl. Surf. Sci. 317 (2014) 35.
adsorption values were 182.1 mg/g and 174.5 mg/g for NO3 and [36] R.F. Xu, S.J. Xin, X. Zhou, Int. J. Pharm. 438 (2012) 258.
[37] D. Cheng, Y.B. Wen, L.J. Wang, Carbohydr. Polym. 123 (2015) 157.
PO43, respectively, which exhibited potential value in removal of [38] P.R. Chang, P.W. Zheng, B.X. Liu, J. Hazard. Mater. 186 (2011) 2144.
nitrate and phosphate wastes in water. REC/Fe3O4-CTAB exhibited [39] J.B. Liu, J. Hao, F. Cheng, J. Environ. Eng. 141 (2015) 04014055.
facile regeneration and good adsorption capacity in the adsorption [40] D.L. Wu, P.W. Zheng, P.R. Chang, X.F. Ma, Chem. Eng. J. 174 (2011) 489.
[41] W. Weng, V.G. Pol, K. Amine, Adv. Mater. 25 (2013) 1608.
cycles. [42] R. Mukherjee, S. De, J. Membr. Sci. 466 (2014) 281.
[43] S. Kilpimaa, H. Runtti, T. Kangas, U. Lassi, T. Kuokkanen, J. Ind. Eng. Chem. 21
(2015) 1354.
Acknowledgements
[44] A. Keranen, T. Leiviska, O. Hormi, J. Tanskanen, J. Environ. Manage. 147 (2015) 46.
[45] A. Sowmya, S. Meenakshi, Int. J. Biol. Macromol. 64 (2014) 224.
This research was supported by the Science and Technology [46] A. Sowmya, S. Meenakshi, Int. J. Biol. Macromol. 69 (2014) 336.
Project of Jiangxi Provincial Office of Education (KJLD12082); the [47] L. Deng, Z. Shi, J. Alloys Compd. 637 (2015) 188.
[48] J.B. Lu, D.F. Liu, J. Hao, Chem. Eng. Res. Des. 93 (2015) 652.
Fundamental Research Funds for the Central Universities [49] X. Liu, L.F. Zhang, Powder Technol. 277 (2015) 112.
(DC201502080402), the Scientific Research Fund of Liaoning [50] S.M. Prabhu, S. Meenakshi, Powder Technol. 268 (2014) 306.