Separation and Purification Technology 178 (2017) 40–48

Contents lists available at ScienceDirect

Separation and Purification Technology

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

Synthesis of magnetite/non-oxidative graphene composites and their

application for arsenic removal
Yeojoon Yoon a,b, Mengliang Zheng a, Yong-Tae Ahn a, Won Kyu Park b,c, Woo Seok Yang b,⇑,
Joon-Wun Kang a,⇑
a
Department of Environmental Engineering, Yonsei University, Yonseidae-gil 1, Wonju-si, Gangwon-do 26493, Republic of Korea
b
Electronic Materials and Device Research Center, Korea Electronics Technology Institute (KETI), 25 Saenari-ro, Bundang-gu, Seongnam-si, Gyeonggi-do 13509, Republic of Korea
c
School of Advanced Materials Science and Engineering, Sungkyunkwan University, 2066 Seobu-ro, Jangan-gu, Suwon-si, Gyeonggi-do 16419, Republic of Korea

a r t i c l e i n f o a b s t r a c t

Article history: Since graphene-based materials have been investigated to adsorb many kinds of contaminants such as
Received 31 October 2016 heavy metals especially arsenic, the fabrication costs of them are critically important for their application
Received in revised form 6 January 2017 to the practical adsorption process. Here, we fabricate a non-oxidative graphene with mass production
Accepted 9 January 2017
and synthesis magnetite/non-oxidative graphene (M-nOG) composites for arsenic removal. The M-nOG
Available online 12 January 2017
showed great capacity for adsorption of arsenic against its low material cost. We found that the arsenite
was more influenced by surface complexation and the arsenate was favorable for intraparticle diffusion
Keywords:
in the adsorption process using M-nOG. To confirm the feasibility of M-nOG as adsorbents for arsenic
Arsenic
Non-oxidative graphene
removal, the effect of various conditions such as pH, temperature, competing anions, and humic acid
Magnetite on the arsenic removal were evaluated. Moreover, the repetitive reuse and regeneration of M-nOG were
Adsorption performed. In conclusion, the M-nOG was cost-effective and feasible to use practically as adsorbents for
Regeneration arsenic removal.
Ó 2017 Elsevier B.V. All rights reserved.

1. Introduction
The occurrence of arsenic in groundwater has been issued
around the world due to its toxicity and the potential for chronic
exposure via drinking water [1–3]. Arsenic contamination affects
millions of people especially in Bangladesh, Vietnam, Cambodia,
India, southwest USA, Canada, Argentina, Chile, Mexico, and neigh-
boring countries and results in long-term exposure to arsenic in
drinking water that can cause skin, lungs, liver, kidney, and bladder
cancer [4–6]. The World Health Organization and the United States
Environmental Protection Agency have been adopted the guideline
of 10 lg/L of arsenic as the drinking water standard [7,8]. Various
studies have investigated arsenic remediation using different tech-
nologies, including co-precipitation/coagulation [9], ion exchange
[10], adsorption [11], and membrane filtration [12,13]. Among
these possible treatment processes, adsorption is considered to
be easier and safer technology to handle as compared to the con-
taminated sludge produced by co-precipitation/coagulation, more
⇑ Corresponding authors.
E-mail addresses: wsyang@keti.re.kr (W.S. Yang), jwk@yonsei.ac.kr (J.-W. Kang).
versatile than ion exchange, and less expensive than membrane fil-
tration [14].
Since graphene, a new generation of nano-carbon material, was
discovered [15], the graphene-based material has recently been
evaluated to remove arsenic from water in adsorption process
[16–18]. Although those studies reported that the adsorption using
the graphene-based material was considered as a promising tech-
nology for arsenic removal due to the outstanding removal effi-
ciency, if the unit cost of adsorbent production is expensive, the
advantages of adsorption process would be nonsensical. Mostly,
graphene oxide (GO) and reduced graphene oxide (rGO) have been
used to synthesize the graphene-based adsorbents for arsenic
removal [16,17,19]. Even though the GO has a lot of oxygen-
containing functional groups which can lead to synthesize with
more arsenic adsorption supporting compounds such as iron oxi-
des [19], the general fabrication method of GO, Hummers method
[20], spend large amount of chemicals and long time of reactions.
Moreover, the effort to fabricate a nearly pristine graphene such
as rGO [21], the reduction product of GO, is not essential in terms
of the adsorbent for water treatment process.
Thus, here, we report on the fabrication of non-oxidative gra-
phene (nOG) which is a few-layer graphene for mass production
at a lower cost and in a shorter time than GO and rGO. In this work,

http://dx.doi.org/10.1016/j.seppur.2017.01.025
1383-5866/Ó 2017 Elsevier B.V. All rights reserved.
 Y. Yoon et al. / Separation and Purification Technology 178 (2017) 40–48
Nomenclature
As(III) arsenite
As(V) arsenate
qt amount of arsenic adsorbed at time t (mg/g)
qe equilibrium adsorption capacity (mg/g)
qe(exp) experimental qe value (mg/g)
qe(cal) calculated qe value by the kinetic model (mg/g)
kf pseudo-first-order rate constant (min1)
ks pseudo-second-order rate constant (g/mg min)
kid intraparticle diffusion rate constant (mg/g min1/2)
C intraparticle diffusion constant related to the thickness
of the boundary layer.
R2 correlation coefficient
QL Langmuir constant reflecting the maximum adsorption
capacity (mg/g)
the nOG has been used to synthesize magnetite/nOG composites
(M-nOG) for application to the arsenic removal. The M-nOG has
been characterized by several analytical techniques. The explora-
tory application of this reasonable graphene-based nanocompos-
ites on the removal of arsenite [As(III)] and arsenate [As(V)] in
aqueous solutions is investigated. The effects of the parameters
including adsorption time, initial arsenic concentration, solution
pH and temperature, and competing anions and humic acid on
the arsenic removal properties are studied. Besides, repetitive
reuse and regeneration of the M-nOG are performed. This study
highlighted the facile applicability of this reasonable material in
environmental pollution cleanup.
2. Experimental
2.1. Materials and chemicals
Natural graphite powder (99%) was obtained from Samjung
CNG Co. Ltd. (Cheongdo, Korea, Rep.). Ammonium sulfate ((NH4)2-
SO4), ammonium hydroxide (NH4OH), sulfuric acid (H2SO4, 95%),
iron(II) chloride tetrahydrate (FeCl2
4H2O), and iron(III) chloride
hexahydrate (FeCl3
6H2O) were purchased from Sinopharm Chem-
ical Reagent Co. Ltd. (Shanghai, PR China). As(III) and As(V) solu-
tions were prepared using NaAsO2 (Fluka, Zwijndrecht,
Netherlands) and Na2HAsO4
H2O (Sigma-Aldrich, St. Louis, USA),
respectively. In the experiment for an effect of coexisting anions
and natural organic matter (NOM) on arsenic adsorption, chloride
(Cl) nitrate (NO  
3 ), bicarbonate (HCO3 ), fluoride (F ), sulfate
(SO2
4 ), phosphate (PO 3
4 ), and humic acid (IHSS Suwannee River
humic acid standards) were used. These anions were purchased
from Sigma-Aldrich (St. Louis, USA) and distilled-deionized (DDI)
water was used in all experiments.
2.2. Preparation of nOG and M-nOG composite
nOG was synthesized from the natural graphite and there were
no oxidation or reduction processes using chemicals. For a basic
recipe for the nOG, 3 g of (NH4)2SO4 were added to 100 mL of
95% H2SO4, and then 1 g of graphite was mixed and stirred with
the solution at 50 °C for 30 min to expand the structure of graphite.
The mixture was washed with DDI water for several times until to
reach pH 7, and then it was dried at 100 °C for 24 h. Finally, nOG
was obtained by physical exfoliation method using thermal treat-
ment at 600 °C for 1 h. In the scale-up manufacturing process,
4 kg of nOG could be produced on one production line.
41
b Langmuir constant related to the free energy of adsorp-
tion (L/mg)
Ce equilibrium arsenic concentration after adsorption (mg/
L)
KF Freundlich constant related to the adsorption capacity
((mg/g)(L/mg)1/n)
1/n Freundlich isotherm exponent reflecting the adsorption
intensity
Qs sips constant reflecting the maximum adsorption capac-
ity (mg/g)
Ks sips isotherm constant (L/mg)
s sips isotherm exponent
For M-nOG, 100 mg of nGO was dispersed in 100 mL of DDI
water by ultrasonication for 1 h. A solution of melted 0.2 g of FeCl2-

4H2O in 5 mL of 0.5 M HCl and melted 0.54 g of FeCl3
6H2O in
10 mL of DDI water was prepared and mixed. The mixed solution
was slowly added to the nGO solution while stirring under a nitro-
gen atmosphere for 15 min. Then, 16 mL of 30% NH4OH solution
was quickly added to the solution to regulate the pH to 10. After
stirring for 45 min, the solution was filtered and washed using
ethanol and water several times. Finally, after the powder was
dried in vacuum oven at 70 °C for 24 h, M-nOG was obtained.
To analyze the characteristics of the nOG and M-nOG, a field
emission scanning electron microscope (FE-SEM, JSM-7600F, JEOL),
Energy Dispersive X-ray Spectroscopy (EDS, JSM-7600F, JEOL),
transmission electron microscopy (TEM, JEM-3010, JEOL), X-ray
photoelectron spectroscopy (XPS, VG Microtech ESCA2000, JEOL),
Raman (Bruker FRA 160/S, BRUKER), and X-ray diffraction (XRD,
D8 ADVANCE, BRUKER) were used.
2.3. Adsorption studies
All of the adsorption experiments were performed in capped
glass vials with 30 mL of arsenic solution and 3 mg of adsorbent
(M-nOG). All of the samples were shaken for a specific contact time
in a shaking incubator (SI-300, Lab Companion). The pH of the
solutions was adjusted using 0.1 M NaOH or HCl to the desired ini-
tial pH values. All samples were filtered using 0.2 lm syringe filter
(Chromafil CA-20, MACHEREY-NAGEL) after reaction, and the con-
centration of residual arsenic was analyzed using an inductively
coupled plasma-atomic emission spectrometer (ICP/IRIS, Thermo
Jarrell Ash Co., USA). The amount of arsenic adsorbed (qe, mg/g)
was calculated from the following Eq. (1)
ðC 0  C e ÞV
qe ¼ ð1Þ
m
where C0 and Ce are the initial and equilibrium concentration of
arsenic in solution (mg/L), V is the volume of solution (L), and m
is the mass of the M-nOG composite (g).
3. Results and discussion
3.1. Characterization of M-nOG
Several characterization techniques were used to determine the
properties of M-nOG and to confirm the proper synthesis of mag-
netite onto nOG. XPS spectra indicate the chemical states of ele-
ments, especially those of carbon (C1s), oxygen (O1s), and iron
42 Y. Yoon et al. / Separation and Purification Technology 178 (2017) 40–48
(Fe2p) in this study (Fig. 1A–C). The C1s spectrum of M-nOG shows
that approximately 70% of the carbon was not oxidized (CAC,
284.5 eV), 19% participated in CAO bonds (representing hydroxyl
and epoxide groups, 286.4 eV), 7% participated in C@O bonds (car-
bonyl, 287.6 eV), and 4% participated in OAC@O bonds (carboxyl,
288.9 eV) (Fig. 1A). A relatively low fraction of oxygen within func-
tional groups was observed because no oxidative chemicals were
used in the fabrication of M-nOG. The O1 s spectrum of M-nOG
showed significant peaks at both 530.4 and 532.3 eV resulting from
the contribution of oxygen to the magnetite composite (Fig. 1B).
Additionally, peaks corresponding to Fe2p3/2 and Fe2p1/2 were
detected at 711.3 and 724.8 eV (Fig. 1C), respectively, indicating
that magnetite was successfully synthesized on nOG.
The Raman spectrum of M-nOG shows three significant peaks
(Fig. 1D). One peak is a G peak at 1590 cm1, which signifies
the sp2 carbon atoms in a graphitic 2D hexagonal lattice. The D
peak at 1350 cm1 indicates the sp3 carbon atoms of defects
and disorder [22]. The G peak of M-nGO is higher than the D peak
because no chemical oxidation or reduction processes occur,
meaning that this method results in fewer defects (carbon lattice
distortion) in M-nGO. The clear shape of the 2D peak at
2700 cm1 indicates that graphite characteristics remain, as a
2D peak is always observed in graphite samples [23]. The 2D peak
is also sensitive to the stacking order of the graphene sheets and
the number of layers [24]: the remarkable 2D peak indicates that
the M-nOG composite is composed of a few randomly arranged
layers of graphene flakes.
The XRD patterns and SEM images also indicated that magnetite
was synthesized onto nOG (Fig. 1E). The intense diffraction peaks
observed with XRD indexed to planes (2 2 0), (3 1 1), (4 0 0), (4 2 2),
(5 1 1), and (4 4 0), which correspond to magnetite in M-nGO. Fur-
thermore, the crystallite size of nOG was obtained by the XRD peak
data containing full width at half maximum, Bragg angle, and
wavelength (EVA program, Fig. S1). From this value of 128 Å
(angstrom), the number of layers of nOG could be estimated at
Fig. 1. XPS spectra of M–nGO: (A) C1 s spectra, (B) O1 s spectra, and (C) Fe2p spectra; (D) Raman spectra and (E) XRD patterns of M-nOG; (F) SEM images and EDS data of nOG
and M-nOG.
approximately 38 layers because the interplanar spacing of
graphite is 0.34 nm.
Fig. 1F shows typical SEM images and EDS values for nOG and
M-nOG. Upon comparing the SEM images of nOG and M-nOG,
the matrix of magnetite particles and some wrinkles can be
observed on the M-nOG surface. The EDS data clearly confirm the
existence of iron (Fe) on the M-nOG surface. The TEM and EDS
mapping images also showed different surface features between
nOG and M-nOG, especially the presence or absence of iron
(Fig. S2). Closer examination using TEM analysis reveals that the
average size of the magnetite nanoparticles is approximately 5–
9 nm. The specific surface area of the M-nOG was estimated using
the BET method and it was 189.94 m2/g.
3.2. Adsorption kinetics
A kinetic study of As(III) and As(V) adsorption on M-nOG was
carried out to understand the adsorption properties of arsenic on
M-nOG. The adsorption of As(III) and As(V) onto the M-nOG as a
function of contact time at pH 7 is shown in Fig. 2. The kinetic pro-
files show a rapid uptake stage within the initial 6 h of contact time
followed by slower uptake thereafter until equilibrium is reached
for both the As(III) and As(V) cases. In subsequent results, 24 h
was used as the equilibration time for As(III) and As(V) on the
M-nOG. To determine the adsorption mechanism from the experi-
mental kinetic results, pseudo-first-order (Fig. S3 and Eq. (S2)),
pseudo-second-order, and Weber and Morris intraparticle diffu-
sion models were investigated in this study.
The pseudo-second-order model has been used to describe
adsorption kinetics [25,26]. This model not only considers the
adsorbate in solution but also the adsorbent in the adsorption pro-
cess. Thus, the equation can be derived from a second-order for-
mula (Eq. (2)).
t 1 1
¼ þ
t ð2Þ
qt ks q2e qe
 Y. Yoon et al. / Separation and Purification Technology 178 (2017) 40–48
Fig. 2. Effect of contact time on adsorption of As(III) and As(V) by M-nOG (M-nOG
dose = 0.1 g/L, arsenic concentration = 1 mg/L, pH 7, temperature = 25 °C).
where qt (mg/g) is the amount of arsenic adsorbed at time t (min),
qe (mg/g) is the adsorption amount at the equilibrium time, and ks
(g/mg min) is the pseudo-second-order rate constant. The plot fea-
tures of t/qt versus t for As(III) and As(V) are shown in Fig. 3A. The qe
and ks values could be calculated from the intercept and slope of the
plots. The plots were found to be linear with good correlation coef-
ficients (R2) close to 1.0, and the calculated qe(cal) value was in good
agreement with experimentally determined qe(exp) values rather
than values determined using a pseudo-first-order model (Table 1).
This result suggests that arsenic adsorption is controlled by the
availability of M-nOG surface sites rather than by arsenic concen-
tration (adsorbate concentration is assumed to be controlling in
the pseudo-first-order model) [27]. The kinetic adsorption of
arsenic on M-nOG can be defined more favorably by the pseudo-
second-order model.
The adsorption of adsorbate from solution by the adsorbent is
generally driven by four steps of mass transfer [28]: (i) diffusion
of adsorbate to the film surrounding the adsorbent from the bulk
space, (ii) diffusion from the film to the adsorbent surface, (iii)
intraparticle diffusion from the surface to adsorbing sites, and
(iv) adsorption onto the adsorbent sites. Of these steps, (ii) film dif-
fusion and (iii) intraparticle diffusion are the major mechanisms
Fig. 3. Adsorption kinetic modeling of As(III) and As(V) on M-nOG: (A) Pseudo-second-order model plots; (B) Intraparticle diffusion model plots (M-nOG dose = 0.1 g/L,
arsenic concentration = 1 mg/L, pH 7, temperature = 25 °C).
43
that determine the rate of adsorption [29]. Thus, the intraparticle
diffusion model proposed by Weber and Morris [30] was used to
describe the adsorption mechanism of arsenic on M-nOG (Eq. (3)).
qt ¼ kid t1=2 þ C ð3Þ
where kid is the intraparticle diffusion rate constant and C is indica-
tive of the intercept, which is related to the thickness of the bound-
ary layer. Plots of adsorbed As(III) and As(V) per unit mass of M-nOG
(qt) versus t1/2 are presented in Fig. 3B. The adsorption reaction is
only controlled by intraparticle diffusion if the Weber and Morris
plot shows a straight line. For our results, however, the data exhib-
ited dual-linear plots consisting of a first linear part (phase I) and a
second straight line (phase II). This behavior suggests that intraparti-
cle diffusion is not the only rate-limiting step in the adsorption pro-
cess. These two distinct phases could be attributed to different rates
of mass transfer during the initial and final adsorption steps. Phase I,
which exhibits a much higher slope with rapid arsenic uptake by
M-nOG, is controlled by intraparticle diffusion and involves immedi-
ate arsenic bonding at the most readily available adsorbing sites on
the M-nOG surface. Thus, as shown in Fig. S2L, the arsenic could be
detected on the M-nOG after arsenic adsorption process. On the
other hand, phase II might be due to slow diffusion of the adsorbate
from surface sites into inner pores or be governed by a surface com-
plexation of adsorption reaction [31,32].
The intraparticle diffusion rate constants of As(III) and As(V) for
phase I (kid,1) and II (kid,2) obtained from the slopes of linear plots
are listed in Table 1 along with R2 values. The kid,1 value of As(V)
was similar to or slightly higher than that of As(III), but the kid,2
value of As(III) was approximately three times higher than that
of As(V). These phenomena indicated that intraparticle diffusion
was more favorable during As(V) adsorption compared to As(III),
while the adsorption of As(III) could be influenced by surface com-
plexation more than that of As(V).
3.3. Adsorption isotherms
The adsorption equilibrium of arsenic was evaluated with dif-
ferent initial As(III) and As(V) concentrations to determine the
adsorption capacity of M-nOG for arsenic. It is clear from Fig. 4 that
the adsorption first increased rapidly and then increased slowly
with increasing arsenic concentration due to the saturation of
M-nOG sites available for arsenic adsorption. Adsorption capacities
of 38 mg/g for As(III) and 14 mg/g for As(V) were observed for
44 Y. Yoon et al. / Separation and Purification Technology 178 (2017) 40–48

Table 1
Kinetic parameters, experimental qe(exp), and calculated qe(cal) values for the adsorption of As(III) and As(V) by M-nOG.

Pseudo-first-order model qe(exp) (mg/g) kf  102 (min1) qe(cal) (mg/g) R2

As(III) 4.99 0.32 3.42 0.936
As(V) 3.76 0.39 1.98 0.872
Pseudo-second-order model qe(exp) (mg/g) ks  102 (min1) qe(cal) (mg/g) R2
As(III) 4.99 0.31 5.18 0.998
As(V) 3.76 0.98 3.84 0.999
Intraparticle diffusion model kid,1 (mg/g min1/2) R2 kid,2 (mg/g min1/2) R2
As(III) 0.1574 0.976 0.0409 0.9186
As(V) 0.1912 0.984 0.0123 0.8144

Fig. 4. Adsorption isotherms of arsenic on M-nOG; (A) As(III), (B) As(V) (M-nOG dose = 0.1 g/L, pH 7, temperature = 25 °C, adsorption time = 24 h).

M-nOG at 25 °C and pH 7. These values were much higher than
those previously reported for magnetic nanoparticle adsorbents
such as magnetic multi-granular nanoclusters (2.4 mg As(V)/g)
[33], hematite coated magnetite particles (2.1 mg As(V)/g) [34],
commercial nanomagnetite (0.209 mg As(III)/g and 0.205 mg As
(V)/g) [35,36], and 20 nm size nanomagnetite (29.2 mg As(III)/g
and 5.9 mg As(V)/g) [37]. Even though other graphene-based mate-
rials such as M-GO and M-rGO have been reported to have shown
greater arsenic adsorption capacity than M-nOG of this work, a
performance/cost ratio should be concerned for commercialization
when determining whether graphene-based materials could be
used in their applications as a feasible adsorbent. As shown in
Table 2, since the nOG used in this study was only 2 USD/g, the
value of its arsenic adsorption capacity/cost ratio was much higher
than those of M-GO and M-rGO. It means that the M-nOG was fea-
sible to use in the application for arsenic adsorption in terms of
material cost.
Three adsorption isotherm models (the Freundlich [38], Lang-
muir [39], and Sips [40] models) were used in this study to
describe the experimental adsorption data (Fig. 4). The equations
of the isotherm models were presented below: Langmuir (Eq.
(4)), Freundlich (Eq. (5)), and Sips (Eq. (6)). The symbols and coef-
ficients used in the equations are defined in the Nomenclature
section.
Q L bC e
qe ¼ ð4Þ
1 þ bC e
qe ¼ K F ðC e Þ1=n ð5Þ
Q s K S C se
qe ¼ ð6Þ
1 þ K S C se
Two-parameter isotherm models such as the Langmuir and Fre-
undlich models are typically used to study the adsorption proper-
ties of arsenic [11,16,17]. As shown in Fig. 4, however, these two
models did not accurately fit the experimental results in this study.
The Langmuir isotherm reveals that adsorption takes place at
specific homogeneous sites of the adsorbent. It means that once
an adsorbate adheres to an adsorbing site, no further adsorption
can occur at the same site [41]. This behavior leads to a constant
monolayer adsorption capacity at high adsorbate concentrations.
Alternatively, at low adsorbate concentrations, the Langmuir iso-
therm effectively reduces to a linear isotherm and thus obeys

Table 2
Comparative evaluation of adsorption capacity and performance/cost ratio among the graphene-based materials such as M-GO, M-rGO, and M-nOG (materials cost: GO = 100
USD/g, rGO = 150 USD/g, and nOG = 2 USD/g; same amount of magnetite was used to synthesize all the graphene-based materials).

Adsorbents Experimental conditions Adsorption capacity Ref. Adsorption

(mg/g) capacity/cost ratio
pH m/V (g/L) Temp (°C) RT (h) As(III) As(V) As(III) As(V)
M-GO 7 0.1 25 24 85 37 [19] (previous study) 1.7 0.74
M-rGO 7 0.1 25 24 57 12 [19] (previous study) 0.76 0.16
M-nOG 7 0.1 25 24 38 14 This study 38.0 14.0
 Y. Yoon et al. / Separation and Purification Technology 178 (2017) 40–48 45

Henry’s law [42]. Conversely, the Freundlich isotherm model can

be applied to nonideal adsorption on heterogeneous surfaces as
well as multilayer adsorption. According to our results (Fig. 4),
the Freundlich model was favorable at low arsenic concentrations.
However, the adsorption trend showed no additional adsorption
similar to the Langmuir model at high arsenic concentrations.
Therefore, the three-parameter Sips model, which is also known
as the combined Langmuir-Freundlich isotherm, was used to fit
to the experimental data, yielding a better fit compared to the indi-
vidual Langmuir and Freundlich models.
The Sips isotherm model was developed to enhance the perfor-
mance of the two-parameter isotherm models [43]. The isotherm
parameters derived from each isotherm model are summarized
in Table 3. Of the three models investigated, the Sips isotherm
model best fit the experimental data, as clearly shown by R2 values
over 0.990 for both As(III) and As(V). The maximum adsorption
capacities calculated by the Sips model (Qs) were also closer to
the experimental results (38 mg/g for As(III) and 14 mg/g for As
(V) than the capacities calculated by the Langmuir model (QL). Fig. 5. Effect of pH on arsenic adsorption by M-nOG (M-nOG dose = 0.1 g/L, arsenic
concentration = 5 mg/L, pH 4, 5, 6, 7, 8, 9, 10, temperature = 25 °C, adsorption
time = 24 h).

3.4. Effect of pH
a wide pH range. A higher As(III) adsorption capacity was shown
The adsorption of As(III) and As(V) on M-nOG was investigated
particularly over a pH range of 6–9 than below pH 6 (positively
at pHs ranging from 4 to 10 (Fig. 5). Arsenic speciation is signifi-
charged surface) or over pH 9 (negatively charged surface). This
cantly affected by pH. As(III) species exist in nonionic form
behavior suggests that the adsorption mechanism of As(III) on
(H3AsO3) in most pH values conditions below pH 9.2
M-nGO is surface complexation rather than an electrostatic inter-
(9.2 < pH < 12.1: H2AsO 2
3 , 12.1 < pH < 13.4: HAsO3 , pH > 13.4:
action. These pH effect results might confirm why As(III) is more
AsO33 ), while As(V) species are present in negative ionic form
preferably adsorbed onto M-nOG than As(V) at neutral pH.
under most pH conditions above pH 2.2 (pH < 2.2: H3AsO4,
2.2 < pH < 6.9: H2AsO 2 3
4 , 6.9 < pH < 11.5: HAsO4 , pH > 11.5: AsO4 )
[44]. The pH also affects the surface charge of the adsorbent parti- 3.5. Effect of temperature
cles. The pH dependence of arsenic adsorption on M-nOG could be
explained by the point of zero charge (pHPZC) of the adsorbent. The To understand the effect of temperature on the adsorption of
surface of the adsorbent is positively charged when the equilib- arsenic, experiments were performed at 5, 15, 25, 35, and 45 °C
rium pH is below pHPZC, whereas the surface is negatively charged for the equilibrium time (24 h); the results are shown in Fig. 6.
when the equilibrium pH is above pHPZC. Iron oxides such as mag- Arsenic removal increased when the temperature was increased
netite typically have pHPZC values in the range 7–9 [45]. From our from 5 to 35 °C, while further increases in temperature resulted
study, the pHPZC value of M-nOG was determined to be pH 7.1. in decreased arsenic removal. These phenomena are consistent
As the pH is increased, the adsorption of As(V) on M-nGO with the results from Pokhrel et al. [48] who observed an increase
decreases because of the negatively charged surface sites on the in arsenic removal from 5 to 30 °C and Mondal et al. [49] who
adsorbent and increased competition between hydroxide ions reported a decrease in arsenic removal from 30 to 60 °C.
(OH) on As(V). This result indicates that the electrostatic interac- The reason for the increase in arsenic removal efficiency with
tion between the positively charged surface of M-nOG and anionic increased temperature in the range of 5–35 °C was not known,
As(V) species is a major factor for As(V) removal. Raven et al. [46]
and Guo et al. [47], who investigated the adsorption of As(V) on
ferrihydrite and cellulose loaded with iron oxyhydroxide respec-
tively, also reported similar results. As(III) exists as a neutral form
species below pH 9.2, so repulsive reactions between As(III) species
and the adsorbent are weak for most pH values. As shown in the
results (Fig. 5), As(III) was favored for adsorption by M-nOG over

Table 3
Isotherm parameters of Langmuir, Freundlich, and Sips models for As(III) and As(V)
adsorption by M-nOG.

Isotherm type Isotherm parameters As(III) As(V)

Langmuir QL (mg/g) 24.6 9.44
b (L/mg) 0.60 0.78
R2 0.856 0.767
Freundlich KF (mg /g)(L/mg)1/n 7.49 3.82
1/n 0.29 0.21
R2 0.974 0.981
Sips Qs (mg/g) 40.6 16.4
KS 0.12 0.25
Fig. 6. Effect of temperature on arsenic adsorption by M-nOG (M-nOG dose = 0.1 g/
s 0.78 0.44
L, arsenic concentration = 5 mg/L, pH 7, temperature = 5, 15, 25, 35, and 45 °C,
R2 0.995 0.994
adsorption time = 24 h).
46 Y. Yoon et al. / Separation and Purification Technology 178 (2017) 40–48
but it was found that this was evident by the high heat of adsorp-
tion related to thermodynamic mechanism [48]. On the other
hand, the decrease in arsenic removal efficiency with increased
temperature over 35 °C agrees with the exothermic nature of
adsorption process. It seems that the intraparticle diffusion process
has less control over the adsorption of As(III) and As(V) as the
diffusion-controlled adsorption is endothermic in nature [50].
The mobility of ions could increase when the temperature
increases, as a result surface precipitation could decrease. In other
words, with the increase in temperature the stability of the bonds
between the active sites of M-nOG and arsenic moiety decreases.
Thus, the adsorption capacity decreases with the increase in tem-
perature [49]. The temperature for maximum arsenic adsorption
was determined to be 35 °C in this study. This result was similar
to the result obtained by Chandra et al. [16] who reported 30 °C
as the temperature for maximum arsenic adsorption.
3.6. Effects of competing anions and humic acid
In natural water sources, especially groundwater, several inor-
ganic anions and NOM might coexist with arsenic species. These
inorganic anions and organic matter could compete with arsenic
for available adsorbing sites during the adsorption process. Thus,
competitive reactions for environmentally major anionic compo-
nents such as Cl, NO   2
3 , HCO3 , F , SO4 , PO4
3
and organic com-
pounds such as humic acid on M-nOG were investigated. The
effects of competing anions and humic acid on arsenic removal
are shown in Fig. 7. In this study, a pH of 7 was chosen, as the
pH of natural water is generally near neutral. The adsorption
capacities of As(III) and As(V) were 11.3 mg/g and 5.5 mg/g,
respectively, for the control with an initial arsenic concentration
of 5 mg/L.
From Fig. 7A, it can be seen that the presence of Cl, NO 3 , and
HCO 3 only slightly inhibited As(III) and As(V) adsorption (94–
98% compared to the control), while the effect of SO2 4 on arsenic
adsorption was relatively high (65–72%). PO34 was found to signif-
icantly decrease both As(III) (31%) and As(V) (52%) adsorption on
M-nOG. These phenomena could be caused by different sorption
mechanisms for the competing anions. Cl, NO 
3 , and HCO3 are just
adsorbed as diffuse ions on the outer surfaces of the adsorbent
while SO24 form outer-sphere complexes with strong electrostatic
interactions [51–53]. Furthermore, arsenic and PO3 4 can form
inner-sphere complexes via a ligand substitution reaction on the
adsorbent surface because they have similar atomic structures
and chemical properties [1,54]. Therefore, PO3 4 adsorbs strongly
onto iron oxide materials as well as arsenic [55–57], allowing
Fig. 7. Effect of coexisting (A) anions and (B) humic acid on arsenic adsorption by M-nOG (M-nOG dose = 0.1 g/L, arsenic concentration = 5 mg/L, pH 7, temperature = 25 °C,
adsorption time = 24 h, anions concentration = 1 mM, humic acid concentration = 1, 3, 5 mg/L as TOC).
the anion to effectively compete with arsenic for M-nOG surface
sites.
For humic acid (Fig. 7B), even though the inhibition of both As
(III) and As(V) adsorption increased with increased humic acid con-
centration, As(III) was affected more than As(V). Because ligand
exchange between the carboxyl/hydroxyl functional groups of
humic acid and iron oxide surfaces was the dominant interaction
mechanism, [58] As(III), which has a strong surface complexation
mechanism, was inhibited to a greater extent than As(V) by humic
acid. Nevertheless, more As(III) is adsorbed onto M-nOG than As(V)
even with 5 mg/L of humic acid as TOC (total organic carbon).
Moreover, there might be another mechanism for the inhibitory
effects on removal of both As(III) and As(V) by the formation of
as-NOM complexes [59].
3.7. Repetitive reuse and regeneration
For practical use as an adsorbent, the number of reuse cycles
and the regeneration of the material are very important for opti-
mizing the adsorption process and reducing operational costs. In
this study, five repetitive reuses to adsorb arsenic followed by
regeneration procedures were performed for 24 h using 1.0 M
NaOH as a desorption solution. After regeneration, the re-
adsorption of arsenic onto regenerated M-nOG was also investi-
gated; the results are shown in Fig. 8.
Following cyclic reuse of M-nOG, it was found that both As(III)
and As(V) were continuously removed even after the fifth cycle,
although the adsorption capacity was decreased. These samples
were assumed to not reach equilibrium adsorption due to the
unsaturation of available adsorbing sites in M-nOG for arsenic.
The total adsorbed arsenic over five cycles was 32.7 mg/g and
13.1 mg/g for As(III) and As(V), respectively, which is less than
the maximum adsorption capacity of As(III) (38 mg/g) and As(V)
(14 mg/g). To achieve a significant capacity again, a regeneration
process should be required. Before the regeneration experiment
was performed, neither As(III) nor As(V) were detected in the elu-
ent (data not shown) when distilled water served as the desorption
agent. This result suggests that arsenic-loaded M-nOG is stable in
water, Furthermore, As(III) or As(V) adsorbed onto M-nOG are
not easily released in a water environment under normal condi-
tions. The remaining arsenic could tightly bind to the adsorbent
effective sites and can be categorized as non-hazardous. Con-
versely, when the NaOH solution was used as the desorption agent,
As(III) and As(V) were released and the desorption ratios for As(III)-
loaded and As(V)-loaded M-nOG were 73.4% and 68.8%, respec-
tively. Although the regenerated adsorbent can be used again for
 Y. Yoon et al. / Separation and Purification Technology 178 (2017) 40–48
Fig. 8. Repetitive reuses and re-adsorption after regeneration of M-nOG for arsenic removal (M-nOG dose = 0.1 g/L, pH 7, temperature = 25 °C, adsorption time = 24 h, initial
arsenic concentration of each cycle and re-adsorption test = 5 mg/L, regeneration: Using 1.0 M NaOH for 24 h).
re-adsorption, it probably cannot retain the same removal effi-
ciency due to the loss of effective adsorbent sites [60]. In this study,
a significant (94.5% for As(III) and 92.4% for As(V) compared to the
adsorption capacity of the first cycle) arsenic re-adsorption capac-
ity after regeneration was observed, even after five consecutive
reuses. These results confirm that M-nOG can be used as an adsor-
bent of arsenic more than once following desorption by NaOH
solution.
4. Conclusions
We have produced the nOG in bulk by a physical exfoliation
method using a thermal treatment. Even though the results of this
study indicated that M-nOG showed lower removal efficiency of
arsenic than other graphene-based materials such as M-GO and
M-rGO, it showed much higher efficiency than other magnetite
composite adsorbents. Moreover, considering the ratio of arsenic
adsorption capacity/material cost, M-nOG was much more efficient
adsorbent for arsenic than both M-GO and M-rGO. The kinetic
study of As(III) and As(V) adsorption on M-nOG indicated that
the pseudo-second-order model was the most favorable. The intra-
particle diffusion was more favorable during As(V) adsorption
compared to As(III), while the adsorption of As(III) could be influ-
enced by surface complexation more than that of As(V). In the
study of adsorption isotherm models, the Sips isotherm represents
a better fit with the experimental data than does the Langmuir and
Freundlich model. This means that the arsenic adsorption mecha-
nism of M-nOG is suitable for the combined Langmuir-Freundlich
isotherm. As the pH is increased, the adsorption of As(V) on M-
nGO decreases because of the negatively charged surface sites on
the adsorbent and increased competition between OH on As(V).
On the other hand, the higher As(III) adsorption capacity was
shown over a pH range of 6–8, which were close to the pHPZC
(7.1) of M-nOG. The temperature for maximum both As(III) and
As(V) adsorption was determined to be 35 °C. According to the
experiment for the effects of competing anions on arsenic removal,
PO3
4 was found to significantly decrease arsenic adsorption on M-
nOG. In terms of the regeneration of M-nOG, a significant arsenic
readsorption capacity after the regeneration was observed even
in the five consecutive reuses of the M-nOG. The results confirm
that M-nOG can be applied as a good adsorbent of arsenic more
than once after the desorption process by the NaOH solution.
47
Acknowledgements
This work is supported by the National Research Foundation of
South Korea (NRF) grant funded by the Korea Government (MSIP)
(2016R1A2B4015598) and Energy Efficiency & Resources of the
Korea Institute of Energy Technology Evaluation and Planning
(KETEP) grant funded by the South Korea Government Ministry
of Knowledge Economy (20142010102690).
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.seppur.2017.01.
025.
References
[1] P.L. Smedley, D.G. Kinniburgh, A review of the source, behaviour and
distribution of arsenic in natural waters, Appl. Geochem. 17 (2002) 517–568.
[2] M. Bissen, F.H. Frimmel, Arsenic – a review. Part 1: Occurrence, toxicity,
speciation, mobility, Acta Hydrochim. Hydrobiol. 31 (2003) 9–18.
[3] K. Ohno, T. Yanase, Y. Matsuo, T. Kimura, M.H. Rahman, Y. Magara, Y. Matsui,
Arsenic intake via water and food by a population living in an arsenic-affected
area of Bangladesh, Sci. Total Environ. 381 (2007) 68–76.
[4] W.R. Cullen, K.J. Reimer, Arsenic speciation in the environment, Chem. Rev. 89
(1989) 713–764.
[5] M.F. Hughes, Arsenic toxicity and potential mechanisms of action, Toxicol. Lett.
133 (2002) 1–16.
[6] B.K. Mandal, K.T. Suzuki, Arsenic round the world: a review, Talanta 58 (2002)
201–235.
[7] WHO, Guidelines for Drinking Water Quality, WHO (World Health
Organization), Geneva, 2011.
[8] USEPA, Federal Register, vol. 66, USEPA, 2001, pp. 6976–7066.
[9] R.C. Cheng, S. Liang, H.-C. Wang, M.D. Beuhler, Enhanced coagulation for
arsenic removal, J. Am. Water Works Assoc. 86 (1994) 79–90.
[10] D.A. Clifford, G. Ghurye, A.R. Tripp, Arsenic removal from drinking water using
ion-exchange with spent brine recycling, J. Am. Water Works Assoc. 35 (2003)
119–130.
[11] D. Mohan, C.U. Pittman, Arsenic removal from water/wastewater using
adsorbents – a critical review, J. Hazard. Mater. 142 (2007) 1–53.
[12] P. Brandhuber, G. Amy, Arsenic removal by a charged ultrafiltration membrane
influences of membrane operating conditions and water quality on arsenic
rejection, Desalination 140 (2001) 1–14.
[13] Y. Yoon, Y. Hwang, M. Ji, B.H. Jeon, J.W. Kang, Ozone/membrane hybrid process
for arsenic removal in iron-containing water, Desalin. Water Treat. 31 (2011)
138–143.
[14] J.L. Schnoor, Environmental Modeling: Fate and Transport of Pollutants in
Water, Air, and Soil, John Wiley & Sons, New York, 1996.
48 Y. Yoon et al. / Separation and Purification Technology 178 (2017) 40–48
[15] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos, I.V.
Grigorieva, A.A. Firsov, Electric field effect in atomically thin carbon films,
Science 306 (2004) 666–669.
[16] V. Chandra, J. Park, Y. Chun, J.W. Lee, I.C. Hwang, K.S. Kim, Water-dispersible
magnetite-reduced graphene oxide composites for arsenic removal, ACS Nano
4 (2010) 3979–3986.
[17] G.D. Sheng, Y.M. Li, X. Yang, X.M. Ren, S.T. Yang, J. Hu, X.K. Wang, Efficient
removal of arsenate by versatile magnetic graphene oxide composites, RSC
Adv. 2 (2012) 12400–12407.
[18] B. Paul, V. Parashar, A. Mishra, Graphene in the Fe3O4 nano-composite
switching the negative influence of humic acid coating into an enhancing
effect in the removal of arsenic from water, Environ. Sci.: Water Res. Technol. 1
(2015) 77–83.
[19] Y. Yoon, W.K. Park, T.M. Hwang, D.H. Yoon, W.S. Yang, J.W. Kang, Comparative
evaluation of magnetite-graphene oxide and magnetite-reduced graphene
oxide composite for As(III) and As(V) removal, J. Hazard. Mater. 304 (2016)
196–204.
[20] W.S. Hummers, R.E. Offeman, Preparation of graphitic oxide, J. Am. Chem. Soc.
80 (1958) 1339.
[21] C. Gomez-Navarro, J.C. Meyer, R.S. Sundaram, A. Chuvilin, S. Kurasch, M.
Burghard, K. Kern, U. Kaiser, Atomic structure of reduced graphene oxide, Nano
Lett. 10 (2010) 1144–1148.
[22] A.C. Ferrari, J.C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S.
Piscanec, D. Jiang, K.S. Novoselov, S. Roth, A.K. Geim, Raman spectrum of
graphene and graphene layers, Phys. Rev. Lett. 97 (2006).
[23] R.P. Vidano, D.B. Fischbach, L.J. Willis, T.M. Loehr, Observation of raman band
shifting with excitation wavelength for carbons and graphites, Solid State
Commun. 39 (1981) 341–344.
[24] A.C. Ferrari, J. Robertson, Interpretation of Raman spectra of disordered and
amorphous carbon, Phys. Rev. B 61 (2000) 14095–14107.
[25] Y.S. Ho, G. McKay, Pseudo-second order model for sorption processes, Process
Biochem. 34 (1999) 451–465.
[26] Y.S. Ho, Review of second-order models for adsorption systems, J. Hazard.
Mater. 136 (2006) 681–689.
[27] Y. Liu, New insights into pseudo-second-order kinetic equation for adsorption,
Colloid Surf. A 320 (2008) 275–278.
[28] M. Badruzzaman, P. Westerhoff, D.R.U. Knappe, Intraparticle diffusion and
adsorption of arsenate onto granular ferric hydroxide (GFH), Water Res. 38
(2004) 4002–4012.
[29] A. Bhatnagar, Y. Choi, Y. Yoon, Y. Shin, B.H. Jeon, J.W. Kang, Bromate removal
from water by granular ferric hydroxide (GFH), J. Hazard. Mater. 170 (2009)
134–140.
[30] W.J. Weber Jr., J.C. Morris, Kinetics of adsorption on carbon from solution, J.
Sanit. Eng. Div. Am. Soc. Civil Eng. 89 (1963) 31–60.
[31] J.C. Igwe, A.A. Abia, C.A. Ibeh, Adsorption kinetics and intrap articulate
diffusivities of Hg, As and Pb ions on unmodified and thiolated coconut fiber,
Int. J. Environ. Sci. Technol. 5 (2008) 83–92.
[32] Z. Velickovic, G.D. Vukovic, A.D. Marinkovic, M.S. Moldovan, A.A. Peric-Grujic,
P.S. Uskokovic, M.D. Ristic, Adsorption of arsenate on iron(III) oxide coated
ethylenediamine functionalized multiwall carbon nanotubes, Chem. Eng. J.
181 (2012) 174–181.
[33] S.H. Lee, J. Cha, K. Sim, J.K. Lee, Efficient removal of arsenic using magnetic
multi-granule nanoclusters, Bull. Korean Chem. Soc. 35 (2014) 605–609.
[34] K. Simeonidis, T. Gkinis, S. Tresintsi, C. Martinez-Boubeta, G. Vourlias, I.
Tsiaoussis, G. Stavropoulos, M. Mitrakas, M. Angelakeris, Magnetic separation
of hematite-coated Fe3O4 particles used as arsenic adsorbents, Chem. Eng. J.
168 (2011) 1008–1015.
[35] T. Turk, I. Alp, H. Deveci, Adsorptive removal of arsenite from water using
nanomagnetite, Desalin. Water Treat. 24 (2010) 302–307.
[36] T. Turk, I. Alp, H. Deveci, Adsorption of As(V) from water using nanomagnetite,
J. Environ. Eng. 136 (2010) 399–404.
[37] S. Yean, L. Cong, C.T. Yavuz, J.T. Mayo, W.W. Yu, A.T. Kan, V.L. Colvin, M.B.
Tomson, Effect of magnetite particle size on adsorption and desorption of
arsenite and arsenate, J. Mater. Res. 20 (2005) 3255–3264.
[38] H.M.F. Freundlich, Over the adsorption in solution, J. Phys. Chem. 57 (1906)
385–470.
[39] I. Langmuir, The adsorption of gases on plane surface of glass, mica and
platinum, J. Am. Chem. Soc. 40 (1916) 1361–1368.
[40] R. Sips, Combined form of Langmuir and Freundlich equations, J. Chem. Phys.
16 (1948) 490–495.
[41] S. Kundu, A.K. Gupta, Arsenic adsorption onto iron oxide-coated cement
(IOCC): regression analysis of equilibrium data with several isotherm models
and their optimization, Chem. Eng. J. 122 (2006) 93–106.
[42] Y.S. Ho, J.F. Porter, G. Mckay, Equilibrium isotherm studies for the sorption of
divalent metal ions onto peat: copper, nickel and lead single component
systems, Water Air Soil Poll. 141 (2002) 1–33.
[43] S.M. Maliyekkal, L. Philip, T. Pradeep, As(III) removal from drinking water
using manganese oxide-coated-alumina: performance evaluation and
mechanistic details of surface binding, Chem. Eng. J. 153 (2009) 101–107.
[44] E. Lacasa, P. Canizares, M.A. Rodrigo, F.J. Fernandez, Electro-oxidation of As(III)
with dimensionally-stable and conductive-diamond anodes, J. Hazard. Mater.
203 (2012) 22–28.
[45] I.A. Katsoyiannis, A.I. Zouboulis, Removal of arsenic from contaminated water
sources by sorption onto iron-oxide-coated polymeric materials, Water Res. 36
(2002) 5141–5155.
[46] K.P. Raven, A. Jain, R.H. Loeppert, Arsenite and arsenate adsorption on
ferrihydrite: kinetics, equilibrium, and adsorption envelopes, Environ. Sci.
Technol. 32 (1998) 344–349.
[47] X.J. Guo, F.H. Chen, Removal of arsenic by bead cellulose loaded with iron
oxyhydroxide from groundwater, Environ. Sci. Technol. 39 (2005) 6808–6818.
[48] D. Pokhrel, T. Viraraghavan, Arsenic removal from an aqueous solution by
modified A-niger biomass: batch kinetic and isotherm studies, J. Hazard.
Mater. 150 (2008) 818–825.
[49] P. Mondal, C. Balomajumder, B. Mohanty, A laboratory study for the treatment
of arsenic, iron, and manganese bearing ground water using Fe3+ impregnated
activated carbon: effects of shaking time, pH and temperature, J. Hazard.
Mater. 144 (2007) 420–426.
[50] V.C. Srivastava, M.M. Swamy, I.D. Mall, B. Prasad, I.M. Mishra, Adsorptive
removal of phenol by bagasse fly ash and activated carbon: equilibrium,
kinetics and thermodynamics, Colloid Surf. A 272 (2006) 89–104.
[51] L. Sigg, K.U. Goss, S. Haderlein, H. Harms, S.J. Hug, C. Ludwig, Sorption
phenomena at environmental solid surfaces, Chimia 51 (1997) 893–899.
[52] K.F. Hayes, G. Redden, W. Ela, J.O. Leckie, Surface complexation models – an
evaluation of model parameter-estimation using fiteql and oxide mineral
titration data, J. Colloid Interf. Sci. 142 (1991) 448–469.
[53] S.J. Hug, In situ Fourier transform infrared measurements of sulfate adsorption
on hematite in aqueous solutions, J. Colloid. Interf. Sci. 188 (1997) 415–422.
[54] K.D. Kwon, J.D. Kubicki, Molecular orbital theory study on surface complex
structures of phosphates to iron hydroxides: calculation of vibrational
frequencies and adsorption energies, Langmuir 20 (2004) 9249–9254.
[55] B.A. Manning, S. Goldberg, Modeling competitive adsorption of arsenate with
phosphate and molybdate on oxide minerals, Soil Sci. Soc. Am. J. 60 (1996)
121–131.
[56] A. Jain, R.H. Loeppert, Effect of competing anions on the adsorption of arsenate
and arsenite by ferrihydrite, J. Environ. Qual. 29 (2000) 1422–1430.
[57] Y. Gao, A. Mucci, Acid base reactions, phosphate and arsenate complexation,
and their competitive adsorption at the surface of goethite in 0.7 M NaCl
solution, Geochim. Cosmochim. Ac. 65 (2001) 2361–2378.
[58] B.H. Gu, J. Schmitt, Z.H. Chen, L.Y. Liang, J.F. Mccarthy, Adsorption and
desorption of natural organic-matter on iron-oxide – mechanisms and models,
Environ. Sci. Technol. 28 (1994) 38–46.
[59] I. Ko, J.Y. Kim, K.W. Kim, Arsenic speciation and sorption kinetics in the As-
hematite-humic acid system, Colloid Surf. A 234 (2004) 43–50.
[60] S.K. Maji, Y.H. Kao, C.W. Liu, Arsenic removal from real arsenic-bearing
groundwater by adsorption on iron-oxide-coated natural rock (IOCNR),
Desalination 280 (2011) 72–79.