Research Article

www.acsami.org

Rapid Removal and Separation of Iron(II) and Manganese(II) from

Micropolluted Water Using Magnetic Graphene Oxide
Han Yan, Haijiang Li, Xue Tao, Kun Li, Hu Yang,* Aimin Li, Shoujun Xiao, and Rongshi Cheng
State Key Laboratory of Pollution Control and Resource Reuse, School of the Environment, School of Chemistry and Chemical
Engineering, Nanjing University, Nanjing 210093, PR China
*
S Supporting Information

ABSTRACT: A novel two-dimensional carbon-based magnetic nanomaterial,

magnetic graphene oxide (MGO), was prepared and then used as an ecient
adsorbent. MGO showed rapid and complete removal of iron(II) (Fe) and
manganese(II) (Mn) from micropolluted water bodies over a wide pH range.
After saturated adsorption, MGO could be rapidly separated from water under
an external magnetic eld. Results of the adsorption equilibrium study
indicated that the adsorption of Fe and Mn by MGO took place via
monolayer heterogeneous and spontaneous processes resulting from the
heterogeneity of the MGO surface as well as from the electrostatic
interactions between surface acidic groups of MGO and metal ions. In
addition, both the Fe and Mn uptake of MGO was very slightly aected by
NaCl, although it decreased with increased humic acid in solutions. In an Fe/
Mn binary aqueous system, both metal ions can be eciently removed at low
concentrations, but MGO showed preferential adsorption of Fe in a
concentrated aqueous mixture. The adsorption behavior in the binary system was due to dierent anities of surface oxygen-
containing functional groups on MGO to Fe and Mn. Finally, unlike traditional approaches in recycling and reusing an adsorbent,
the Fe- and Mn-loaded MGO can be directly applied as a new adsorbent to achieve the ecient removal of uoride from aqueous
solutions.
KEYWORDS: magnetic graphene oxide, removal of iron(II) and manganese(II), selective adsorption, adsorption mechanism

1. INTRODUCTION relatively simple operation.8 Many kinds of ecient adsorbents

Groundwater containing iron(II) (Fe) and manganese(II) have been reported.912 Among these, the application of
(Mn) is widely distributed in China and coexists with other graphene oxide (GO), a two-dimensional (2D) carbon-based
types of contaminants in dierent water bodies. The amount of nanomaterial, has attracted signicant attention because of its
Fe in Chinese groundwater typically is below 510 mg L1 and advantages, such as a large surface area, more activated
that of Mn ranges from 0.5 to 2.0 mg L1, respectively.1 It has functionalized sites, easy preparation, and good biocompati-
been reported that the presence of excessive Fe and Mn in bility.9,1315 These features ensure that GO can facilitate the
water results in an unpleasant color and odor as well as laundry rapid and ecient removal of pollutants, and numerous studies
staining.2 Their presence also increases health risks, such as on the application of GO in water treatment and environmental
cognitive disorder in human beings, especially after long-term remediation have been performed in recent years.16,17 More-
intake of water rich in Fe or Mn.3,4 When Fe- or Mn-rich water over, the rapid separation of adsorbents after saturated
is employed in paper and textile production, printing, and adsorption from water is also signicant for recycling both
dyeing as well as in the chemical and ne leather industries, the sorbates and adsorbents. However, this may become dicult if
quality of the nal products is reduced signicantly.2 Therefore, the adsorbent has a very small size. Recently, magnetic
ecient removal of Fe and Mn has always been a major issue in adsorbents have proven to be easily and rapidly separated
water treatment. The maximum Fe and Mn levels in drinking from liquid media under an external magnetic eld.18
water in China are 0.3 and 0.1 mg L1, respectively, (GB5749- In the current work, a novel GO-based magnetic nanoma-
2006) which are strictly monitored.57 terial, magnetic graphene oxide (MGO), was prepared after
Fe and Mn are conventionally removed by aeration and being impregnated with magnetism. We then carried out the
oxidation combined with ltration or sedimentation.5,6 adsorptive treatment for the removal of Fe and Mn using MGO
Although these techniques are commonly used, they are also as the adsorbent. Fe and Mn contents in synthetic wastewater
time-consuming; thus, they are not ideal for rapid purication.
Adsorption is capable of removing trace levels of pollutants, Received: April 22, 2014
including Fe and Mn, and has gained popularity in water Accepted: April 30, 2014
treatment applications because of its high eciency and

XXXX American Chemical Society A dx.doi.org/10.1021/am502377n | ACS Appl. Mater. Interfaces XXXX, XXX, XXXXXX
ACS Applied Materials & Interfaces Research Article

were controlled at low levels as micropolluted water bodies in Mn aqueous solutions were prepared from FeCl24H2O and MnCl2
order to simulate actual Chinese groundwater sources. To 4H2O, respectively.
explore the adsorption mechanism, the fundamental adsorption 2.3.1. Eect of Initial Solution pH. The eects of the initial solution
pH on the removal of Fe and Mn at 298 K were tested. The initial pH
behaviors of MGO, including the eects of the initial solution
ranges were 4.0 to 7.5 for Fe and 4.0 to 10.0 for Mn. The lower and
pH, additives, isothermal adsorption equilibrium, and adsorp- upper pH limits were set to avoid both the dissolution of magnetic
tion kinetics, were investigated. Furthermore, because the two Fe3O4 particles under acidic conditions and the precipitation of metal
metal ions often coexist in water, a study on the adsorptive hydroxides in alkaline media. Two sets of initial Fe and Mn
removal of both Fe and Mn in a binary aqueous system was also concentrations were designed: one was 0.04 mmol L1 (2.24 mg
done. Finally, a novel method was carried out to reuse the Fe- L1 for Fe and 2.20 mg L1 for Mn) for the simulation of real Chinese
and Mn-loaded adsorbents, in which they were directly applied groundwater sources as micropolluted water bodies and the other was
as new adsorbents for the removal of another pollutant. relatively high, 1.5 mmol L1 (84.0 mg L1 for Fe and 82.5 mg L1 for
Mn), for the comprehensive investigation of the MGO adsorption
Fluoride (F), a common pollutant found in various water
performance.
sources, was selected as the target pollutant for secondary About 0.03 g of MGO was dosed and dispersed in 30 mL of a Fe or
adsorption. Mn aqueous solution with dierent initial pH levels. The mixture was
constantly agitated for 4 h to achieve adsorption equilibrium. The nal
2. MATERIALS AND METHODS concentrations of Fe and Mn were determined on a Thermo M6
atomic absorption spectrometer. Both ame and furnace methods
2.1. Materials. Graphite was purchased from Tianjin Institute of were adopted, and proper calibration was also performed to ensure the
Chemical Reagents. H 2 SO 4 , HCl, KMnO 4 , NaNO 3 , NaOH, consistency of the dierent methods. All analyses were peformed in
(NH4)2Fe(SO4)26H2O, and humic acid (HA) were obtained from triplicate, and the nal metal concentration results are the average of
Sinopharm Chemical Reagent Co. Ltd. FeCl36H2O, ammonia three runs, for which the relative error was lower than 1.0%.
solution, FeCl24H2O, MnCl24H2O, NaF, and NaCl were purchased In addition, eects of initial solution pH on the adsorption
from Nanjing Chemical Reagent Co. Ltd. All of the reagents were of performance of pure Fe3O4 nanoparticles for the removal of Fe and
analytical grade. Mn in respective single-component water bodies were also studied
2.2. Preparation of MGO. Prior to MGO preparation, GO was under the same conditions as those for MGO for further comparison.
obtained on the basis of Hummers method,19 for which KMnO4 was The initial concentrations of Fe and Mn were 84.0 and 82.5 mg L1,
used as oxidant. Next, the structure of GO was characterized by Raman respectively.
spectroscopy, wide-angle X-ray diraction (XRD), and X-ray photo- Equilibrium adsorption capacity (qe, mg g1) was calculated from
electron spectroscopy (XPS), the details of which have been described the metal concentration change in the adsorption process using the
in our previous work.20 The resulting Raman and XRD spectra of GO following equation
are shown in Supporting Information Figures S1 and S2, respectively.
The average aromatic cluster size in GO, based on the intensity ratio of (C0 Ce)V
qe =
G/D peaks in the Raman spectrum,21 was 1.32 nm. Meanwhile, the m (1)
interplanar distance (d) of the oxidized carbon planes according to the 1
where C0 and Ce (mg L ) are the initial and equilibrium Fe or Mn
Bragg equation was about 0.80 nm from the XRD spectrum, and the
concentrations of the mixture, respectively, V (L) is the total volume
oxygen content on the GO surface estimated from the XPS spectrum of the solution at equilibrium, and m (g) is the dried weight of the
was approximately 32.81%. Furthermore, the content of various adsorbent.
surface oxygen-containing acidic groups of GO (i.e., carboxyl, lactonic, The removal rate (R %) was calculated at the same time according
and phenolic groups), as determined by the Boehm titration method, to the following equation
were 2.3, 4.1, and 0.9 mmol g1, respectively22 (see Supporting
Information Text S1 for details). C0 Ce
R (%) = 100%
The MGO was prepared through in situ coprecipitation of Fe3O4 on C0 (2)
GO. The experimental details were as follows: 1.0 g of GO was
1
dispersed in water together with 2.53 g of (NH4)2Fe(SO4)26H2O and where C0 and Ce (mg L ) are the initial and equilibrium Fe or Mn
1.74 g of FeCl36H2O. A few drops of 0.1 mol L1 HCl were added concentrations of the aqueous mixture, respectively.
into the solution to inhibit Fe(III) hydrolysis. The mixture was heated 2.3.2. Adsorption Equilibrium Study. An adsorption equilibrium
at 353 K under N2 protection. Then, 10 mL of ammonia solution was experiment was also conducted at 298 K at a pH of approximately 5.5.
added, and the mixture was stirred vigorously. The resultant black Then, 0.03 g of MGO was dosed and dispersed in 30 mL of Fe or Mn
suspension was constantly stirred at 353 K for 1 h, after which it was aqueous solutions with dierent initial metal concentrations ranging
ltered and washed. The nal product, MGO, was obtained and kept from 0.1 to 90 mg L1. Each aqueous mixture was agitated for 4 h to
in water for further use. The Fe3O4 content in MGO was 59%, as achieve adsorption equilibrium. A similar analysis method using the
determined from the concentration of dissolved iron ions by an atomic atomic absorption spectrometer, as mentioned above, was employed
absorption spectrometer once a known amount of MGO was to detect the initial and nal metal concentrations. The metal uptakes
equilibrated in a 0.1 mol L1 HCl aqueous solution. Then, the were calculated on the basis of eq 1.
content of GO component was deduced to be approximately 41%. 2.3.3. Adsorption Kinetics Study. An adsorption kinetics study was
Aside from the Raman and XRD spectra shown in Supporting also carried out at 298 K and pH 5.5. About 0.2 g of MGO was
Information Figures S1 and S2, MGO underwent vibration magneto- dispersed in 200 mL of a Fe or Mn solution, and the mixture was then
meter (VSM), transmission electron microscopy (TEM), and zeta stirred continuously. Two sets of initial concentrations of Fe and Mn,
potential characterizations. The VSM spectrum is shown in Supporting 0.04 and 1.5 mmol L1, respectively, were designed using the same
Information Figure S3. The characteristic peaks of GO and Fe3O4 conditions as described in Section 2.3.1. About 1.0 mL of the mixed
were both observed in the Raman and the XRD spectra of MGO; solution was taken out at desired time intervals to analyze the current
moreover, the saturated magnetization of MGO was 20 emu g1, metal concentrations. Meanwhile, the same volume of pure water was
indicating that MGO was obtained successfully. The detailed added into the bulk solution to keep the volume constant. The metal
operational conditions of all instrumental characterizations are uptake of MGO at time ti, q(ti) (mg g1) was calculated using the
shown in Supporting Information Table S1. following equation
2.3. Fe and Mn Removal in Respective Single-Component i1
(C0 Cti)Vo 2 C ti 1Vs
Systems. The adsorption behavior of MGO in the removal of Fe and q(ti) =
Mn in their respective single-component systems was studied. Fe and m (3)

B dx.doi.org/10.1021/am502377n | ACS Appl. Mater. Interfaces XXXX, XXX, XXXXXX

ACS Applied Materials & Interfaces Research Article

where C0 and Cti (mg L1) are the initial concentration of Fe or Mn

and its concentration at time ti, respectively. V0 and Vs (L) are the
volume of the mixed solution and that of the sample solution taken out
each time for concentration analysis, respectively. In this equation, Vs
is equal to 1.0 mL. Finally, m (g) represents the mass of the adsorbent.
2.3.4. Eects of Additives. The adsorption performance of MGO in
the presence of additives (i.e., NaCl and HA) at 298 K and pH 5.5 was
studied. About 0.03 g of MGO was immersed in 30 mL of a metal
solution, which contained dierent amounts of additives. Initial
concentrations of Fe and Mn were 2.24 and 2.20 mg L1, respectively.
A similar analysis method using the atomic absorption spectrometer, as
mentioned above, was employed to detect the nal metal
concentrations once adsorption equilibrium was reached.
2.4. Fe and Mn Removal in Binary Systems. The adsorption
behavior of MGO for the removal of both metal ions from the Fe/Mn
binary aqueous system was carried out at 298 K and pH 5.5. In the
adsorption equilibrium experiment, the experimental conditions were
similar to those used in the Fe or Mn single-component systems
described in Section 2.3.2 except that Fe and Mn coexisted in the
solution.
The conditions of the adsorption kinetics experiment were also
similar to that in the Fe or Mn single-component system described in
Section 2.3.3. The initial concentrations of Fe and Mn in the binary
mixture were 2.24 and 2.20 mg L1, respectively.
2.5. Reuse of MGO. Upon completion of saturated adsorption, the
metal-loaded MGOs, denoted MGO-Mn and MGO-Fe, were directly
reused as new adsorbents for the removal of F from water at 298 K.
The eect of the initial solution pH on F adsorption was investigated.
Then, 0.1 g of Fe- or Mn-loaded MGO was immersed in 100 mL of a
1.0 mmol L1 (19 mg L1) F aqueous solution. The suspension was
constantly agitated for 6 h to reach equilibrium. The nal residual F
concentration was the average of three runs determined using a
Mettler Toledo DX-219 F uorine-selective electrode, for which the
relative error was lower than 2.5%. Then, F uptakes were calculated on
the basis of eq 1.
Figure 1. Eects of initial solution pH on the Fe and Mn removal rate
3. RESULTS AND DISCUSSION and residual concentration by MGO in respective single-component
micropolluted water bodies at 298 K. The initial concentrations of Fe
3.1. Removal of Fe and Mn in Their Respective Single- and Mn were 2.24 and 2.20 mg L1, respectively (a); pH dependence
Component Systems. 3.1.1. Eect of Initial Solution pH. of the zeta potentials of MGO, GO, MGO-Mn, and MGO-Fe (b).
MGO was prepared using in situ coprecipitation of Fe3O4 on
GO. The experimental details, including the structural
information on MGO, are described in the Materials and
Methods. MGO was employed as an adsorbent for the removal
of Fe and Mn from the aqueous solutions. pH level is an
important factor to ensure a good adsorption performance of
ionic-type adsorbents. Thus, in this work, the eect of the initial
solution pH on the adsorption of metal ions from single-
component micropolluted water bodies was studied, and the
results are shown in Figure 1a.
As seen in Figure 1a, most of the removal rates of Fe and Mn
under the measured pH range were higher than 95%, and the
maximal removal rates were almost 100% at the optimal pH.
Furthermore, the residual metal concentrations were mostly
compliant with Chinese standards for drinking water quality
(GB5749-2006), indicating that MGO was an eective
adsorbent for use in the adsorption of Fe and Mn from
micropolluted water. The high eciency of Fe and Mn removal Figure 2. Adsorption isotherms of MGO for the removal of Fe and
was mainly due to electrostatic interactions among the metal Mn in their respective single-component water bodies at 298 K and a
ions and the acidic groups of GO in MGO, because neither the pH of approximately 5.5.
hydrophobic aromatic matrix in the nonoxidation region of GO
nor the surface weak positively charged Fe3O423 can eciently that many oxygen-containing groups with negative charges as
bind to Fe or Mn in water. On the basis of Supporting active adsorption sites exist in MGO, thus exhibiting good
Information Figure S4, the pure Fe3O4 nanoparticles had really anity to cationic species. However, the zeta potentials of
almost no contribution to the adsorption of metal ions from MGO were less negative than those of GO over the whole
aqueous solutions. Moreover, the highly negative zeta measured pH range, resulting from the fact that only 41% of the
potentials of MGO, as shown in Figure 1b, further conrmed GO component was included in the MGO. Furthermore, after
C dx.doi.org/10.1021/am502377n | ACS Appl. Mater. Interfaces XXXX, XXX, XXXXXX
ACS Applied Materials & Interfaces Research Article

Table 1. Isotherms Parameters for the Adsorption of Metal Ions onto MGO at 298 K and a pH of Approximately 5.5
Langmuir Freundlich
metal ions qm,exp (mg g1) qm (mg g1) b (L mg1) r2 Kf n r2
Fe 43.2 30.3 1.86 0.9725 16.1 2.37 0.9873
Mn 16.5 15.0 1.23 0.9480 6.65 2.72 0.9722

Table 2. Adsorption Capacities and Equilibrium Time of Various Adsorbents for the Removal of Fe and Mn
adsorbents qFe (mg g1) qMn (mg g1) equilibrium time (min) refs
MGO 1.0a this work
43.2 16.5 15b
granular activated carbon 3.6 2.5 unknown 11
apatite 57.1124.2 18.5126.0 unknown 12
AC derived from agro-residues 0.8 0.9 15c 27
40c
Albizia procera legume substrate 11 8.6 30 28
natural zeolite 1.1 0.08 60 29
adsorbent coal 15 25 60 30
para-phenol-resin 9.5 100 31
natural zeolitic tu 8.6d 1400d 32
a
The initial concentrations of Fe and Mn were 2.24 and 2.20 mg L1, respectively. bThe initial concentrations of Fe and Mn were 84.0 and 82.5 mg
L1, respectively. cEquilibrium time for Fe and Mn, respectively. dIt was measured at 298 K.

loading with Fe and Mn, the zeta potentials of MGO-Fe and The equilibrium data were also subjected to Langmuir and
MGO-Mn both increased, but they were still negative based on Freundlich isothermal models for further investigation of the
Figure 1b. Interestingly, the zeta potentials increase was adsorption mechanism. The Langmuir and Freundlich models
approximatively proportional to the metal ion uptake, resulting were deduced from dierent theoretic assumptions as
in higher zeta potentials of MGO-Fe than those of MGO-Mn. follows:24,25 the former is based on the assumption of
In addition, Figure 1a also shows the wide adsorption monolayer adsorption on a homogeneous surface,24 whereas
window of pH for the removal of both Fe and Mn. Removal the latter is usually applied in describing heterogeneous
rates, however, decreased slightly when pH decreased or systems.25 The equations are expressed as follows
increased further. Some of the GO anionic groups were
protonized under acidic conditions, resulting in a weakened qmbCe
negative charge and adsorption inhibition, as conrmed by the
qe =
1 + bCe (4)
pH dependence of the zeta potentials of both GO and MGO.
By contrast, excessive hydroxide ions competed with GO to
bind with metal ions under alkaline conditions. qe = KCe1/ n (5)
Furthermore, the eect of initial solution pH on Fe and Mn
adsorption from synthetic water with high metal concentrations where qe is the amount of Fe or Mn adsorbed at equilibrium
was also determined in order to give a comprehensive (mg g1), Ce is the concentration of Fe or Mn at equilibrium
investigation of the adsorption performance of MGO (mg L1), qm is the adsorption capacity when the adsorbent is
(Supporting Information Figure S5). From Figures S5 and fully covered (mg g1), b is the Langmuir adsorption constant
1a, it was found that there is a similar variation trend of (L mg1), Kf is the Freundlich isotherm constant, and n
adsorption performance with pH at both low and high metal (dimensionless) is the heterogeneity factor.
concentrations. The tting results are listed in Table 1. The correlation
3.1.2. Adsorption Equilibrium Study. The equilibrium coecients (r2) of the linear forms for the Langmuir and
isotherm has a fundamental role in describing the interactive Freundlich models were both high, with those of the Langmuir
behavior between solutes and adsorbent. Thus, the adsorption model being slightly lower. The result indicated that the
isotherms of MGO for the removal of Fe and Mn in their Langmuir and Freundlich mechanisms may be both involved in
respective single-component systems were measured over a MGO adsorption of Fe and Mn. As mentioned earlier, both Fe
wide metal concentration range (Figure 2). In the beginning, and Mn were adsorbed through electrostatic interactions with
the Fe and Mn uptake increased linearly with the initial metal the acidic groups of GO. Thus, on one hand, the MGO
concentrations. Furthermore, both Fe and Mn were almost adsorption of metal ions was monolayer and partially followed
completely removed by MGO when the initial concentrations the Langmuir model. On the other hand, various oxygen-
were lower than 3.5 mg L1 based on the residual metal containing groups (i.e., carboxyl, lactonic, and phenolic groups)
concentrations. This result indicated that the adsorption sites existed and were distributed on the surface of MGO according
on MGO were sucient and that the adsorption capacity fully to its structural characteristics. The various acidic groups clearly
relied on the amount of Fe or Mn transported from the bulk bear dierent activity and adsorption energy levels,25 thus
solutions to the adsorbent surfaces at lower initial concen- resulting in a heterogeneous surface of the adsorbent that
trations. Meanwhile, at higher initial concentrations, the followed the Freundlich adsorption mechanism. Summarily, the
adsorption sites on the MGO surfaces were saturated, and adsorption of Fe and Mn on MGO was a monolayer
the adsorption of metal ions achieved equilibrium. heterogeneous process.
D dx.doi.org/10.1021/am502377n | ACS Appl. Mater. Interfaces XXXX, XXX, XXXXXX
ACS Applied Materials & Interfaces Research Article

Figure 3. Adsorption kinetics of MGO for removal of Fe and Mn in their respective single-component micropolluted water bodies at 298 K and a
pH of approximately 5.5. The initial concentrations of Fe and Mn were 2.24 and 2.20 mg L1, respectively (a). TEM image of MGO (b). Separation
processes of GO by gravity after 12 h and MGO by magnetism (1200 G) after 8.0 min from their aqueous suspensions after adsorption of metal ions
(c).

In addition, the thermodynamic parameter of G (kJ mol1) and approximately 28.6 and 27.5 kJ mol1, respectively.
for the adsorption process can be calculated according to G = This indicated that the adsorption processes were spontaneous.
RT ln KL. Where R (J mol1 K1) is the gas constant, T (K) is Moreover, the maximum saturated Fe uptake of MGO
the temperature, and KL (L mol1) is the Langmuir equilibrium (about 43.2 mg g1) was evidently higher than the Mn uptake
constant, which can be obtained by multiplying the constant of (about 16.5 mg g1) (Figure 2). This may be attributed to the
b by the molar weight of the metal ions. The G values for the fact that Fe is more polarizable than Mn for the relatively inert
adsorption of Fe and Mn by MGO at 298 K were both negative half-lled electronic arrangement of Mn (3d5) compared with
E dx.doi.org/10.1021/am502377n | ACS Appl. Mater. Interfaces XXXX, XXX, XXXXXX
ACS Applied Materials & Interfaces Research Article

Furthermore, in comparison with other reported adsorb-

ents11,12,2732 (Table 2), MGO had a much shorter equilibrium
time during the adsorption of both Fe and Mn, indicating a
higher removal eciency in terms of both adsorption capacity
and speed. The rapid adsorption may be ascribed to two facts:
stronger anity of metal ions to the acidic groups of GO in
MGO and good dispersibility of the few-layered MGO sheets.
The latter has been conrmed by direct observation under
TEM (Figure 3b), resulting in larger surface areas and more
activated functionalized sites that are also benecial to the rapid
adsorption of pollutants.
To discuss the adsorption mechanisms further, pseudo-rst
and pseudo-second order kinetic models34,35 were applied to
the experimental data. The nonlinear forms of the afore-
mentioned kinetics models are presented as eqs 6 and 7,
respectively

qt = qe qeek1t (6)

qe
qt = qe
qek 2t + 1 (7)

where qe and qt are the amount of Fe or Mn adsorbed onto

adsorbents (mg g1) at equilibrium and at time t (min),
respectively. k1 (min1) and k2 (g mg1 min1) are the rate
constants of pseudo-rst and pseudo-second order models,
respectively.
The simulation results are listed in Supporting Information
Table S2. The r2 of both kinetic models for Fe and Mn were all
very high (>0.99) in the two sets of water systems with low and
high initial metal concentrations. These results indicated that
there were various binding types with dierent anity strengths
Figure 4. Eect of additives, humic acid (a) and NaCl (b), on Fe and between solutes and adsorbent, which resulted from various
Mn adsorption by MGO in their respective single-component
activities of dierent acidic groups on the GO surface. The
micropolluted water bodies at 298 K and a pH of approximately 5.5.
The initial concentrations of Fe and Mn were 2.24 and 2.20 mg L1, results of the adsorption kinetics are fully consistent with those
respectively. of the isothermal adsorption equilibrium discussed earlier.
Rapid separation of adsorbents after saturated adsorption is
also very signicant. MGO, after evenly complexing with
that of Fe (3d6),26 thus resulting in higher Fe uptake and nanomagnets (Figure 3b), exhibited much faster separation in
stronger anity of Fe to those acidic groups. Furthermore, the less than 8.0 min (Figure 3c), whereas its precursor GO
reaction eciency of the two metals with the total active sites without magnetism hardly settled down even after 12 h. This
on MGO can be roughly estimated on the basis of the contents suggests the great application potential of MGO for fast
of GO acidic groups in MGO, metal ion uptake, and charge removal and separation of Fe and Mn from water bodies.
balance. Reaction eciencies were about 51.5 and 20.0% for Fe 3.1.4. Eects of Additives. Furthermore, because there are
and Mn, respectively, thus indicating dierent activity of the many coexisting solutes in real water, the eects of two
acidic groups and the existence of less active adsorption sites on common additives, HA (organic matter) and NaCl (inorganic
MGO that bound to Mn with more diculty than to Fe. matter), on the adsorption of Fe and Mn were studied in their
The adsorption capacities of MGO were also compared with respective single-component micropolluted water bodies
those of some other reported adsorbents11,12,2732 (Table 2). (Figure 4). HA showed an inhibition eect on the adsorption
We found that the Fe and Mn uptake of MGO was higher than of both Fe and Mn, but the inhibition eect on Fe was more
that of most adsorbents, indicating that MGO can eciently notable (Figure 4a). This may be due to the stronger chelating
remove Fe and Mn from polluted water. ability of Fe to HA with its abundant active functional groups.36
3.1.3. Adsorption Kinetics Study. Adsorption kinetics was However, when the concentration of HA was below 2.0 mg L1,
also studied to establish the time course of Fe and Mn uptake the residual concentrations of Fe and Mn were still lower than
on MGO in their respective single-component systems. Figure the upper limits of GB5749-2006.
3a illustrates the variation in the adsorption capacity trends Compared with organic matter, NaCl showed almost no
versus contact time for micropolluted water. The adsorption impact on the adsorption of the two metal ions in the
equilibrium for both Fe and Mn was achieved within 1.0 min, experimental concentration range (Figure 4b). This result
which was very rapid compared with aeration and oxidation indicated that MGO bears salt tolerance in the removal of Fe
treatment that usually takes hours.33 Even for synthetic water and Mn from micropolluted water, which can be ascribed to the
with high metal concentrations, adsorption equilibrium was stronger interaction of Fe and Mn with MGO that cannot be
reached in 15 min (Supporting Information Figure S6). exchanged by Na+ easily.
F dx.doi.org/10.1021/am502377n | ACS Appl. Mater. Interfaces XXXX, XXX, XXXXXX
ACS Applied Materials & Interfaces Research Article

Figure 5. Adsorption behavior of MGO in an Fe/Mn binary water body at 298 K and a pH of approximately 5.5. Adsorption isotherms (a),
adsorption kinetics, and the initial concentrations of Fe and Mn were 2.24 and 2.20 mg L1, respectively (b). Adsorption selectivity (c). Available
selective adsorption processes (d).

3.2. Fe and Mn Removal in Binary Systems. It is single-component systems (Figure 2), the variation trend of Fe
meaningful to investigate the removal of Fe and Mn under was quite similar, but that of Mn was distinctly dierent. From
coexisting conditions given that the two ions frequently coexist Figure 5a, Mn and Fe uptake both increased linearly with initial
in water bodies. The adsorption isotherms of MGO for the metal concentrations under coexisting conditions. Later, the
removal of Fe and Mn in their binary systems are shown in growth trend of Mn uptake slowed, unlike Fe, and decreased
Figure 5a. Compared with their isotherms in their respective after reaching a maximum (4.5 mg L1) as the initial metal
G dx.doi.org/10.1021/am502377n | ACS Appl. Mater. Interfaces XXXX, XXX, XXXXXX
ACS Applied Materials & Interfaces Research Article

the recovery of adsorbents by desorption using large amounts

of eluent. This is an ecient process, although it can also lead
to secondary pollution. Given that the surface structure of
MGO after saturated adsorption is distinctly changed but
remains stable enough under certain conditions, the Fe- and
Mn-loaded adsorbents can be directly applied as new
adsorbents for the removal of another pollutant. We attempted
to adsorb F, a common pollutant in various water sources, onto
metal-loaded MGOs. The results detailed in Figure 6 show that
MGO had almost no eect on F removal, whereas the F uptake
of MGOs after loading metal ions increased. This indicates that
the loaded Fe and Mn play key roles in the adsorption of F
from aqueous solutions. Furthermore, MGO-Fe and MGO-Mn
both exhibited pH independence for the removal of F over the
experimental pH range. This nding may be ascribed to F
having a stronger anity to active adsorption sites (i.e., Fe and
Figure 6. pH dependence of the adsorption capacities of two metal- Mn on MGO), which can exclude other coexisting ions.
loaded MGOs and pure MGO for the removal of F at 298 K. The Considering the negative zeta potentials of MGO-Mn and
initial concentration of F was 19.0 mg L1. MGO-Fe, as shown in Figure 1b, the strong anity between
active adsorption sites and F may be a coordination eect. In
addition, the F uptake of MGO-Fe and MGO-Mn was
concentrations further increased. Moreover, the adsorption
approximately 5.0 and 1.0 mg g1, respectively. These values
rates of both Fe and Mn were still ecient in micropolluted
are quite comparable with those of previously reported
coexisting water, as conrmed by adsorption kinetics experi-
adsorbents,3740 suggesting that metal-loaded MGOs are
ment (Figure 5b) wherein both metal ions were almost
qualied as adsorbents for F removal. Moreover, the higher F
completely removed by MGO from the Fe/Mn binary
uptake of MGO-Fe than that of MGO-Mn may be ascribed to
micropolluted water body based on the residual metal
the higher amount of Fe on MGO.
concentrations. In addition, MGO also exhibited very fast
removal of Fe and Mn (i.e., only 1.0 min) under coexisting 4. CONCLUSIONS
conditions.
However, as the concentrations of Fe and Mn went up, This study highlights the applicability of MGO, a novel 2D
adsorption of Fe was preferred, but that of Mn was inhibited. A carbon-based magnetic nanomaterial, in the removal of Fe and
selectivity coecient, , was introduced for a quantitative Mn from water. MGO was capable of the rapid removal and
description in order to study the adsorption selectivity of MGO separation of both metal ions from micropolluted water bodies
toward Fe further. Here, can be calculated on the basis of the over a wide pH range because of the embedded Fe3O4
following equation nanoparticles. Furthermore, the residual concentrations of Fe
and Mn were fully compliant with Chinese standards for
qe,FeCe,Mn drinking water quality (GB5749-2006). The higher removal
Fe
Mn = eciencies of MGO in terms of both adsorption capacity and
Ce,Feqe,Mn (8) speed were due to a stronger anity of the metal ions for the
where Fe
Mn is the selectivity factor of Fe over Mn, qe,Fe and qe,Mn oxygen-containing groups of GO. Furthermore, the adsorption
are the equilibrium adsorption capacities, and Ce,Fe and Ce,Mn of Fe and Mn took place via monolayer heterogeneous and
are the equilibrium concentrations of Fe and Mn, respectively. spontaneous processes for the heterogeneous MGO surface
On the basis of Figure 5a, Fe Mn was plotted against the initial containing various acidic groups. In addition, the Fe and Mn
metal concentrations listed in Figure 5c, which illustrated the uptake of MGO was unaected by NaCl, but it decreased
strong selectivity of MGO for Fe because most of the Fe Mn greatly in the presence of HA, resulting from the higher
values were much higher than 1.0. As mentioned earlier, there chelating eects of metal ions with organic matter. In the Fe/
are various acidic groups (i.e. carboxyl, lactonic, and phenolic Mn binary system, both metal ions can be eciently removed
groups) on MGO, each with their own adsorption activity. at low metal concentrations if there are sucient active
However, the binding constants of Fe with most of the acids adsorption sites. However, at high metal concentrations, Fe was
were higher than those of Mn,26 which resulted in a stronger selectively removed, but the adsorption of Mn was fully
binding anity of Fe to the active adsorption sites and a higher inhibited. The higher selectivity of MGO for Fe was due to the
selectivity of MGO for Fe. The detailed selective adsorption greater anity of Fe to the active adsorption sites. Finally,
process in the Fe/Mn binary system is described in Figure 5d. metal-loaded MGOs can be applied as new adsorbents for the
In micropolluted water, highly active adsorption sites are direct removal of F. This eective approach provides a zero-
sucient for the ecient removal of Fe and Mn together. cost technique for the reuse of adsorbents. Therefore, the 2D
When the metal concentrations increased, Fe preferentially magnetic nanomaterial is expected to be a highly ecient
adsorbent for use in future water treatment applications.

occupied the highly active adsorption sites, and some amounts
of Mn were expelled and replaced. Moreover, because Mn had
a much lower anity to the residual less active sites on MGO, it ASSOCIATED CONTENT
was dicult to bind eectively, thus leading to a decrease in Mn *
S Supporting Information
uptake and eventual maintenance at a very low level. Details of the Boehm titration of GO; details of the structural
3.3. Reuse of MGO. Reusing adsorbents is very important characterizations of the samples; nonlinear simulation results of
in practical applications. The traditional reuse method involves the kinetics data of MGO based on pseudo-rst and pseudo-
H dx.doi.org/10.1021/am502377n | ACS Appl. Mater. Interfaces XXXX, XXX, XXXXXX
ACS Applied Materials & Interfaces Research Article

second order models; Raman spectra of graphite, GO, and (13) Chen, D.; Feng, H.; Li, J. Graphene Oxide: Preparation,
MGO; XRD spectra of graphite, GO, and MGO; VSM Functionalization, and Electrochemical Applications. Chem. Rev. 2012,
spectrum of MGO; eect of initial solution pH on the Fe 112, 60276053.
and Mn uptake of pure Fe3O4 nanoparticles and MGO; and (14) Compton, O. C.; Nguyen, S. T. Graphene Oxide, Highly
Reduced Graphene Oxide, and Graphene: Versatile Building Blocks
adsorption kinetics of MGO for the removal of Fe and Mn for Carbon-Based Materials. Small 2010, 6, 711723.
single-component water bodies. This material is available free of (15) Wang, J.; Tsuzuki, T.; Tang, B.; Hou, X.; Sun, L.; Wang, X.
charge via the Internet at http://pubs.acs.org. Reduced Graphene Oxide/Zno Composite: Reusable Adsorbent for

Pollutant Management. ACS Appl. Mater. Interfaces 2012, 4, 3084

AUTHOR INFORMATION 3090.
(16) Upadhyay, R. K.; Soin, N.; Roy, S. S. Role of Graphene/Metal
Corresponding Author Oxide Composites as Photocatalysts, Adsorbents and Disinfectants in
*E-mail: yanghu@nju.edu.cn; Tel/Fax: 86-25-89680377. Water Treatment: A Review. RSC Adv. 2014, 4, 38233851.
(17) Chowdhury, R.; Balasubramanian, R. Recent Advances in the
Funding
Use of Graphene-Family Nanoadsorbents for Removal of Toxic
This work was supported by the Natural Science Foundation of Pollutants from Wastewater. Adv. Colloid Interface Sci. 2014, 204, 35
China (grant nos. 51378250 and 51073077) and by the Open 56.
Fund from State Key Laboratory of Pollution Control and (18) Rocher, V.; Siaugue, J. M.; Cabuil, V.; Bee, A. Removal of
Resource Reuse of Nanjing University (grant no. Organic Dyes by Magnetic Alginate Beads. Water Res. 2008, 42, 1290
PCRRF11004). 1298.
(19) Hummers, W. S. J.; Offeman, R. E. Preparation of Graphitic
Notes Oxide. J. Am. Chem. Soc. 1958, 80, 1339.
The authors declare no competing nancial interest. (20) Yan, H.; Tao, X.; Yang, Z.; Li, K.; Yang, H.; Li, A.; Cheng, R.

Effects of the Oxidation Degree of Graphene Oxide on the Adsorption

REFERENCES of Methylene Blue. J. Hazard. Mater. 2014, 268, 191198.
(21) Tuinstra, F.; Koening, J.L. Raman Spectrum of Graphite. J.
(1) Zhang, L. S. The Advanced Water Treatment and Reuse Technology; Chem. Phys. 1970, 53, 11261130.
Chemical Industry Press: Beijing, 2009. (22) Boehm, H. P. Some Aspects of the Surface Chemistry of Carbon
(2) Ning, R. Y. Colloidal Iron and Manganese in Water Affecting RO Blacks and Other Carbons. Carbon 1994, 32, 759769.
Operation. Desalin. Water Treat. 2009, 12, 162168. (23) Ahmed, S. M. Studies of the Dissociation of Oxide Surfaces at
(3) Dahshan, H.; Abd-Elall, A. M. M.; Megahed, A. M. Trace Metal the LiquidSolid Interface. Can. J. Chem. 1966, 44, 16631670.
Levels in Water, Fish, and Sediment from River Nile, Egypt: Potential (24) Langmuir, I. The Adsorption of Gases on Plane Surfaces of
Health Risks Assessment. J. Toxicol. Environ. Health 2013, 76, 1183 Glass, Mica and Platinum. J. Am. Chem. Soc. 1918, 40, 13611402.
1187. (25) Freundlich, H. M. F. Uber Die Adsorption in Lasungen. Z. Phys.
(4) Tang, Q.; Liu, G.; Zhou, C.; Zhang, H.; Sun, R. Distribution of Chem. 1906, 57, 385470.
Environmentally Sensitive Elements in Residential Soils Near a Coal- (26) Dean, J. A. Langes Handbook Of Chemistry; Mcgraw-Hill: New
Fired Power Plant: Potential Risks to Ecology and Childrens Health. York, 1999.
Chemosphere 2013, 93, 24732479. (27) Akl, M. A.; Yousef, A. M.; Abd El Nasser, S. Removal of Iron
(5) Korchef, A.; Kerkeni, I.; Amor, M. B.; Galland, S.; Persin, F. Iron and Manganese in Water Samples Using Activated Carbon Derived
Removal from Aqueous Solution by Oxidation, Precipitation and from Local Agro-Residues. J. Chem. Eng. Proc. Technol. 2013, 4, 110.
Ultrafiltration. Desalin. Water Treat. 2009, 9, 18. (28) Kose, T. D.; Gharde, B. D.; Gholse, S. B. Studies on Albizia
(6) Qin, S. Y.; Ma, F.; Huang, P.; Yang, J. X. Fe(II) and Mn(II) procera Legumes for Effective Removal of Fe(II) and Mn(II) from
Removal from Drilled Well Water: A Case Study from a Biological Aqueous Solution. J. Chem. Pharm. Res. 2012, 4, 20212028.
Treatment Unit in Harbin. Desalination 2009, 245, 183193. (29) Shavandi, M. A.; Haddadian, Z.; Ismail, M. H. S.; Abdullah, N.;
(7) Yuce, G.; Alptekin, C. In situ and Laboratory Treatment Tests for Abidin, Z. Z. Removal of Fe(III), Mn(II) and Zn(II) from Palm Oil
Lowering of Excess Manganese and Iron in Drinking Water Sourced Mill Effluent (POME) by Natural Zeolite. J. Taiwan Inst. Chem. Eng.
from RiverGroundwater Interaction. Environ. Earth Sci. 2013, 70, 2012, 43, 750759.
28272837. (30) Vistuba, J. P.; Nagel-Hassemer, M. E.; Lapolli, F. R.; Lobo
(8) Yan, H.; Yang, L.; Yang, Z.; Yang, H.; Li, A.; Cheng, R. Recio, M. A . Simultaneous Adsorption of Iron and Manganese from
Preparation of Chitosan/Poly(acrylic acid) Magnetic Composite Aqueous Solutions Employing an Adsorbent Coal. Environ. Technol.
Microspheres and Applications in the Removal of Copper(II) Ions 2013, 34, 275282.
from Aqueous Solutions. J. Hazard. Mater. 2012, 229, 371380. (31) Kamel, N. H. M.; Sayyah, E. M.; Abdel-Aal, A. A. Removal of
(9) Sun, Y.; Wang, Q.; Chen, C.; Tan, X.; Wang, X. Interaction Lead, Cobalt and Manganese from Aqueous Solutions Using a New
between Eu(III) and Graphene Oxide Nanosheets Investigated by Modified Synthetic Ion Exchanger. Arch. Appl. Sci. Res. 2011, 3, 448
Batch and Extended X-ray Absorption Fine Structure Spectroscopy 464.
and by Modeling Techniques. Environ. Sci. Technol. 2012, 46, 6020 (32) Rajic, N.; Stojakovic, D.; Jevtic, S.; Zabukovec Logar, N.; Kovac,
6027. J.; Kaucic, V. Removal of Aqueous Manganese Using the Natural
(10) Donia, A. M.; Atia, A. A.; Elwakeel, K. Z. Selective Separation of Zeolitic Tuff from the Vranjska Banja Deposit in Serbia. J. Hazard
Mercury(II) Using Magnetic Chitosan Resin Modified with Schiffs Mater. 2009, 172, 14501457.
Base Derived from Thiourea and Glutaraldehyde. J. Hazard. Mater. (33) Serikov, L. V.; Tropina, E. A.; Shiyan, L. N.; Frimmel, F. H.;
2008, 151, 372379. Metreveli, G.; Delay, M. Iron Oxidation in Different Types of
(11) Bin Jusoh, A.; Cheng, W. H.; Low, W. M.; Noraaini, A.; Megat Groundwater of Western Siberia. J. Soils Sediments 2009, 9, 103110.
Mohd Noor, M. J. Study on the Removal of Iron and Manganese in (34) Lagergren, S. About the Theory of So-Called Sorption of
Groundwater by Granular Activated Carbon. Desalination 2005, 182, Soluble Substances. K. Sven. Vetenskapsakad. Handl. 1898, 24, 139.
347353. (35) Ho, Y. S.; Mckay, G. Sorption of Dye from Aqueous Solution by
(12) Oliva, J.; De Pablo, J.; Cortina, J.-L.; Cama, J.; Ayora, C. The Peat. Chem. Eng. J. 1998, 70, 115124.
Use of Apatite II To Remove Divalent Metal Ions Zinc(II), Lead(II), (36) Batchelli, S.; Muller, F. L. L.; Chang, K.-C.; Lee, C.-L. Evidence
Manganese(II) and Iron(II) from Water in Passive Treatment for Strong but Dynamic Iron-Humic Colloidal Associations in Humic-
Systems: Column Experiments. J. Hazard. Mater. 2010, 184, 364374. Rich Coastal Waters. Environ. Sci. Technol. 2010, 44, 84858490.

I dx.doi.org/10.1021/am502377n | ACS Appl. Mater. Interfaces XXXX, XXX, XXXXXX

ACS Applied Materials & Interfaces Research Article

(37) Kemer, B.; Ozdes, D.; Gundogdu, A.; Bulut, V. N.; Duran, C.;
Soylak, M. Removal of Fluoride Ions from Aqueous Solution by Waste
Mud. J. Hazard. Mater. 2009, 168, 888894.
(38) Sivasankar, V.; Ramachandramoorthy, T.; Chandramohan, A.
Fluoride Removal from Water Using Activated and MnO2-Coated
Tamarind Fruit (Tamarindus indica) Shell: Batch and Column Studies.
J. Hazard. Mater. 2010, 177, 719729.
(39) Ghorai, S.; Pant, K. K. Investigations on the Column
Performance of Fluoride Adsorption by Activated Alumina in a
Fixed-Bed. Chem. Eng. J. 2004, 98, 165173.
(40) Alagumuthu, G.; Rajan, M. Equilibrium and Kinetics of
Adsorption of Fluoride onto Zirconium Impregnated Cashew Nut
Shell Carbon. Chem. Eng. J. 2010, 158, 451457.

J dx.doi.org/10.1021/am502377n | ACS Appl. Mater. Interfaces XXXX, XXX, XXXXXX