Accepted Manuscript

Functionalization of activated carbons with magnetic Iron oxide

nanoparticles for removal of copper ions from aqueous solution

Si-Yong Gu, Chien-Te Hsieh, Yasser Ashraf Gandomi, Zhi-Feng

Yang, Lingyun Li, Chun-Chieh Fu, Ruey-Shin Juang

PII: S0167-7322(18)34843-8
DOI: https://doi.org/10.1016/j.molliq.2018.12.018
Reference: MOLLIQ 10094
To appear in: Journal of Molecular Liquids
Received date: 19 September 2018
Revised date: 26 November 2018
Accepted date: 3 December 2018

Please cite this article as: Si-Yong Gu, Chien-Te Hsieh, Yasser Ashraf Gandomi, Zhi-
Feng Yang, Lingyun Li, Chun-Chieh Fu, Ruey-Shin Juang , Functionalization of activated
carbons with magnetic Iron oxide nanoparticles for removal of copper ions from aqueous
solution. Molliq (2018), https://doi.org/10.1016/j.molliq.2018.12.018

This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.
 ACCEPTED MANUSCRIPT

Functionalization of Activated Carbons with Magnetic Iron Oxide

Nanoparticles for Removal of Copper Ions from Aqueous Solution

Si-Yong Gu1, Chien-Te Hsieh2,3,*, Yasser Ashraf Gandomi3, Zhi-Feng Yang4, Lingyun

PT
Li4, Chun-Chieh Fu5, Ruey-Shin Juang5,6,*

1
Fujian Provincial Key Laboratory of Functional Materials and Applications,

RI
SC
Institute of Material Preparation and Applied Technology, School of Materials

Science and Engineering, Xiamen University of Technology, Xiamen 361024,

NU
Fujian, PR China
MA

2
Department of Chemical Engineering and Materials Science, Yuan Ze University,

Taoyuan 32003, Taiwan

E D

3
Department of Mechanical, Aerospace, and Biomedical Engineering, University
PT

of Tennessee, Knoxville, TN 37996, United States

CE

4
College of Materials Science and Engineering, Fuzhou University, Fuzhou
AC

350108, P. R. China

5
Department of Chemical and Materials Engineering, Chang Gung University,

Guishan, Taoyuan 33302, Taiwan

6
Division of Nephrology, Department of Internal Medicine, Chang Gung Memorial

Hospital, Linkou, Taiwan

1
 ACCEPTED MANUSCRIPT

Prepared for Submission to Journal of Molecular Liquids as a full article

September 19, 2018

Revised November 26, 2018

PT
RI
* Corresponding author. Phone number: 1-865-898-0833; E-mail address:

SC
[email protected] (Prof. C.-T. Hsieh) and [email protected]

(Prof. R.-S. Juang)

NU
MA
E D
PT
CE
AC

2
 ACCEPTED MANUSCRIPT

Abstract

Magnetic activated carbon particles (MACP) were fabricated using a facile co-

precipitation method and employed as an adsorbent for the removal of Cu ions from

aqueous solutions. The application of magnetic Fe3O4 nanoparticles significantly

PT
enhanced the adsorption capacity and regeneration efficiency of the MACP adsorbents.

RI
The Fe3O4 magnetite loading onto the carbon matrix is a crucial factor for an efficient

SC
removal of copper ions. The Langmuir, Freundlich, and Dubinin-Radushkevich (D-R)

adsorption models were used to validate the adsorption isotherms. Both the equilibrium
NU
rate constant (Langmuir model) and the adsorption energy (D-R model) show an
MA

increase in the deposition of magnetite nanoparticles. The presence of magnetic Fe3O4

D

nanoparticles further facilitates the regeneration efficiency of Cu-adsorbed

E

intermediates from the MACP adsorbents. Superior magnetic property enabled via the
PT

utilization of Fe3O4 nanoparticles significantly enhanced the separation efficiency and

CE

regeneration of spent MACP adsorbents. Therefore, the design procedure explained in

AC

this work for preparing the MACP adsorbents can be adopted for various environmental

applications dealing with separation processes.

Keywords: Magnetite nanoparticles; Activated carbon; Liquid-phase adsorption;

Copper ions; Composite adsorbent

3
 ACCEPTED MANUSCRIPT

1. Introduction

The effluents carrying high concentration of heavy metals are extremely toxic to

human and aquatic life; resulting in kidney and liver problems and genotoxic

carcinogen, ultimately deteriorating public human health [1,2]. The main source of

PT
heavy metals polluting the drinking water is the industrial wastewater (primarily from

RI
the modern chemical industries focused on fossil fuels, batteries, metal plating,

SC
pesticides, and many others [3]).

Having a density higher than 5 g cm-3 along with an atomic weight in the range
NU
of 63.5 and 200.6 are the major characteristics of heavy metals (e.g. Cu, Cd, and Mg)
MA

[4,5]. The copper ions (Cu2+) are believed to severely poison the drinking water [5].
D

The excessive ingestion of Cu ions causes numerous health issues including an

E

increased blood pressure and respiratory variations (causing serious damages within the
PT

kidney and liver), convulsion, cramps, vomiting, or even death [3]. To alleviate these
CE

undesired consequences, many techniques have been devoted for removing metal ions
AC

from liquid phase including bioremediation, coagulation, lime softening, chemical

precipitation, membrane filtration, solvent extraction, electrolysis, reverse osmosis, ion

exchange and adsorption [6-9].

Among various techniques, the adsorption method is considered a low-cost,

efficient and economical approach capable of removing various contaminants from

4
 ACCEPTED MANUSCRIPT

aqueous solutions. The selection of adsorbents is a crucial factor influencing the

removal/regeneration efficiency and maintenance cost of the system. Several

adsorbents including activated carbon (AC) [10-13], carbon nanotube [14], graphene

oxide [15,16], chitosan [17], carbon cloths [18], and zeolite [19], have been explored

PT
for the adsorption of organic compounds and metal ions in aqueous solutions.

RI
Recently, magnetic activated carbon particles (MACP) have received

SC
considerable attention thanks to high surface area and superior magnetic properties as

an emerging adsorbent. The magnetic properties enable the spent adsorbents to be easily
NU
removed from the aqueous solution along with the pollutants being adsorbed using a
MA

magnetic bar or under a magnetic field [4,20]. In general, filtration and centrifugation
D

are used after adsorption process. They are time-consuming. The blockage of filters and
E

the loss of adsorbents may occur in filtration method. Centrifugation is not an economic
PT

method. The use of magnetic separation method offers an alternative way to solve these
CE

problems. Therefore, the adsorption process renders any additional filtration or

AC

centrifugation unnecessary if the MACP is used for the removal of the pollutants [20],

leading to an easy operation without any further post treatment. However, the

adsorption mechanism and the regeneration efficiency of MACP adsorbents requires a

detailed investigation to establish a correlation between the design parameters to overall

efficiently during the adsorption/regeneration process.

5
 ACCEPTED MANUSCRIPT

In this work, for an efficient removal of Cu2+ ions from aqueous solutions, high-

performance MACP adsorbent has been synthesized and the adsorption process has

been explored in detail. A novel co-precipitation method is adopted to deposit highly-

crystalline Fe2O3 nanoparticles onto or into AC and the resulting MACP has been used

PT
as an adsorbent. The weight loading of as-prepared Fe2O3 nanoparticles has been

RI
engineered as a controlling parameter for the design of MACP adsorbents. The

SC
deposition process of spent MACP adsorbents under a magnetic field along with

adjusting pH value has been carried out to delineate the regeneration efficiency. A series
NU
of adsorption experiments have been conducted and the resulting adsorption isotherms
MA

have been interpreted using different adsorption models (Langmuir, Freundlich, and
D

Dubinin-Radushkevich (D-R)). This work elaborates an excellent scalable adsorption

E

process for industrial applications with a facile operating procedure, high

PT

removal/regeneration performance and reduced cost (using inexpensive MACP

CE

adsorbents).
AC

2. Materials and Methods

2.1 Synthesis of MACP adsorbents

Schematic diagram for the formation of MACP adsorbents has been illustrated

in Figure 1. Commercial AC powders, made from coconut shell, were prepared

6
 ACCEPTED MANUSCRIPT

through a carbonization step followed by an activation procedure using steam as the

activation agent. Prior to deposition of magnetic nanoparticles, AC powders (ca. 10 g)

were chemically oxidized by impregnating the particles within 4 N HNO3 (1 L) at

95°C for two hours. The chemical oxidation process allowed the implantation of

PT
various oxygen functionalities such as carbonyl, carboxylic and hydroxyl groups, onto

RI
the surface of AC powders. The oxidized AC powders (ca. 10 g) were then put into an

SC
iron-containing solution (1 L), consisted of FeCl3·6H2O and FeSO4·7H2O with a molar

ratio of Fe3+: Fe2+ = 2:1, based on our previous studies [21,22]. Through an ionic
NU
exchange process, the Fe2+ and Fe3+ ions were being adsorbed occupying active sites
MA

on the AC surface. The AC slurry was stirred using a magnetic bar, and the ionic
D

exchange process was conducted at ambient temperature for four hours. To determine
E

the optimal ratio, the MACP adsorbents with different AC to Fe3O4 weight ratios (i.e.,
PT

1:0.5, 1:1, and 1:2, labeled as AF1, AF2, and AF3, respectively) were fabricated. The
CE

AC slurry was then heated to 50°C and kept at this temperature under nitrogen
AC

atmosphere for 1 hour. Followingly, NH3·H2O was slowly dropped into the AC slurry

until the pH value of the slurry reached 10, generating MACP adsorbent. Upon

termination of the ammonia titration, a water/ethanol solution was used to completely

wash as-prepared MACP adsorbents. Finally, the composite adsorbents were

dehydrated in a vacuum oven at 80°C overnight.

7
 ACCEPTED MANUSCRIPT

2.2 Characterization of MACP adsorbents

Field-emission scanning electron spectroscope (FE-SEM, JEOL 6701F) was

adopted to assess the microstructural morphology of as-prepared MACP adsorbents.

PT
The crystalline structure of MACP samples was characterized by X-ray diffraction

RI
(XRD) with Cu-Kα radiation, using an automated X-ray diffractometer (Rigaku,

SC
D/MAX 2500). An automated adsorption apparatus (Micromeritics, ASAP 2020) was

used for analyzing the surface morphology of porous carbons using N2 physisorption at
NU
-196°C. Specific surface areas and pore size distributions of carbon samples were
MA

calculated with the application of the Brunauer-Emmett-Teller (BET) and the Barrett-
D

Joyner-Halenda (BJH) equations, respectively. Before any such analysis, the carbon
E

samples were degassed at 110°C in a vacuum at about 10-3 Torr. The chemical
PT

composition of MACP adsorbents was analyzed using X-ray photoelectron

CE

spectroscopy (XPS). The XP spectra were recorded and subsequently collected using
AC

Fison VG ESCA210 spectrometer and Mg-Kα radiation. The deconvolution of XP

spectra was achieved using the non-linear least square algorithm with a symmetric

Gaussian function. The thermal stability of MACP samples was examined with a

thermogravimetric analyzer (TGA, Perkin Elmer Pyris 1 TGA). The TGA analysis was

carried out under a steady air flow with a ramping rate of 10 °C/min within the entire

8
 ACCEPTED MANUSCRIPT

temperature range between 30 and 800°C.

2.3 Adsorption/desorption experiments

The schematic diagram describing the adsorption/desorption cycle of Cu ions on

PT
MACP adsorbents in liquid phase is shown in Figure 1. The aqueous Cu2+ solution was

RI
prepared via diluting CuSO4·5H2O in distilled water at different concentrations, ranging

SC
from 100 to 900 mg/L. To conduct adsorption experiments, MACP adsorbent (ca. 0.1

g) along with 20 mL of the aqueous solution was placed into a 100 mL glass-stoppered
NU
flask. The amount of pristine AC was approximately 0.1‒0.4 g. Based on the experience
MA

gained from our preliminary studies, the adsorption process reaches an equilibrium in
D

about two hours. The spent MACP adsorbent was easily separated from Cu-containing
E

solution using a magnet. The Cu2+ concentration in the residual solutions was analyzed
PT

with a UV/visible spectrophotometer (Shimadzu, Model UV-2550) at 810 nm. By

CE

comparing the light absorbance of the residual solutions against a calibration curve in
AC

an appropriate range of concentrations (10‒1,000 mg/L), the ionic concentrations were

accurately measured. Consequently, the amount of adsorbate captured by MACP

adsorbents (i.e., equilibrium adsorption capacity, qe) was determined. After adsorption

equilibrium, the spent MACP was collected by using neodymium magnet in our case.

The spent MACP adsorbents were chemically regenerated via adding 0.5 M NaOH

9
 ACCEPTED MANUSCRIPT

solution, serving as a regeneration agent, until the pH reached to 11. The

adsorption/desorption cycles were repeated 7 times to assess the regeneration efficiency.

PT
3. Results and Discussion

RI
3.1 Physicochemical characterization of MACP adsorbents

SC
The surface morphology of MACP adsorbents (i.e., AF1 sample) were

characterized by FE-SEM (Figures 2(a) and 2(b)). As illustrated in Figure 2, a large

NU
amount of magnetite nanoparticles has been attached to the surface of AC powder.
MA

According to Figure 2, the AC powders possess an irregular shape with an average

D

particle size of about several micrometers; whereas the magnetite particles tend to form
E

a spherical configuration with an average diameter of 30 nm covering the AC powders.

PT

The XRD patterns of bare AC, oxidized AC, and iron oxide particles coated AC
CE

powders are illustrated in Figure 2(c). As shown in Figure 2, the appearance of a small
AC

lump within the range of 22° to 26° reveals the presence of amorphous carbon.

Considering the iron oxide nanoparticles, two distinct crystalline phases within the

MACP structure (i.e., AF1 sample) can be identified: hematite and magnetite. The

hematite is usually assigned to a crystalline solid with a hexagonal compact structure;

whereas the magnetite appears as a crystalline oxide demonstrating cubic structure of

10
 ACCEPTED MANUSCRIPT

an inverse spinel type possessing two distinct sites with an asymmetric configuration

[20]. The strong diffraction peaks of MACP indexed to (220), (311), (400), (422), (511),

and (440) planes occur at 2θ = 30.1°, 36.3°, 43.4°, 53.9°, 57.2°, and 62.3°, respectively,

corresponding to the cubic phase nanoparticles of Fe3O4 with a face-centered cubic

PT
structure (JCPDS No. 89-3854) [4].

RI
The surface characteristics of bare AC and MACP adsorbents, including BET

SC
surface areas, micropore, and mesopore volumes are summarized and listed in Table 1

(based on the adsorption data associated with N2). Table 1 clearly shows that the bare
NU
AC adsorbent has the largest surface area of about 647 m2 g-1, whereas the decoration
MA

of magnetite nanoparticles significantly reduces the specific surface areas of MACP

D

samples. According to Table 1, the total pore volume of MACP samples has also been
E

reduced consequently since the pores and channels existing within the AC structure are
PT

partially blocked by the Fe3O4 nanoparticles. It is interesting to note that the mesopore
CE

fraction of MACP samples increases with an increased Fe3O4 nanoparticles loading

AC

(AF1 to AF3). This reflects that the magnetic Fe3O4 nanoparticles although cause a

partial pore blockage, but an extra mesoporous structure is also being formed when the

Fe3O4 nanoparticles are added to the structure. This extra pore structure mainly

originates from the inter-aggregation of Fe3O4 particles and the pertinent continuous

magnetite layers, imparting accessible active surface for the adsorption of nitrogen

11
 ACCEPTED MANUSCRIPT

physisorption. For example, the mesopore fraction of AF3 sample reaches as high as

94.9% inducing a larger mean pore size within the MACP sample.

Figure 3(a) shows survey-scan XPS spectra of bare AC, oxidized AC, and MACP

samples, demonstrating the presence of iron oxide in MACP sample. This finding

PT
reflects that the co-precipitation method is capable of coating Fe3O4 magnetite onto the

RI
surface of AC powders, forming the MACP hybrid. As shown in Figure 3(b), we

SC
observe that Fe‒O (binding energy: 530.3 eV), Fe‒O‒C (binding energy: 532.8 eV) and

C=O (binding energy: 531.6 eV) species are major contributors to O 1s peak. It is worth
NU
noting that the appearance of Fe‒O‒C confirms a strong affinity between magnetic
MA

Fe3O4 nanoparticles and carbonaceous materials. The binding energy peaks at ca. 710.5
D

and 724.3 eV correspond to Fe 2p3/2 and Fe 2p1/2, respectively [23], as depicted in

E

Figure 3(c). The inset of Figure 3(c) presents the deconvolution of Fe 2p3/2 peak at about
PT

710.5 eV, confirming the existence of Fe2+ and Fe3+ mixed oxide (i.e., Fe3O4). This
CE

insight further validates the analysis of XRD patterns. The deconvolution of C 1s peak
AC

for MACP adsorbent was also conducted using multiple Gaussian peak functions as

illustrated in Figure 3(d). In Figure 3, several components with pertinent binding

energies can be identified: sp2 carbon at ~284.4 eV, C‒OH at ~285.8 eV, C=O at ~286.9

eV, and HO‒C=O at ~288.4 eV [24,25]. As shown in Figure 3, the hydroxyl group is

the major contributor to the functional group distribution. Therefore, MACP adsorbent

12
 ACCEPTED MANUSCRIPT

possesses a high O/C ratio (i.e., high oxidation level), leading to a hydrophilic surface.

The surface hydrophilicity is believed to be favorable for the adsorption of metal ions

in liquid phase.

The weight loading of Fe3O4 magnetite on the MACP adsorbent was analyzed by

PT
TGA measurement, as shown in Figure 4. The TGA measurements were performed in

RI
air and the maximal weight loss takes place at 550°C. The weight loss is attributed to

SC
gasification of carbon atoms under oxygen-containing atmosphere, i.e., the burn-off of

AC powders. The residual weight can be evaluated to assess the weight loading of
NU
magnetite on MACP adsorbent. The analysis of the residual weight of the samples
MA

revealed the following order: AF3 (65.1%) > AF2 (50.1%) > AF1 (35.6%), which
D

correlates to the sequence of Fe loading during the samples preparation.

E
PT

3.2 Adsorption performance of Cu ions on MACP adsorbents

CE

The adsorption isotherms of Cu ions onto bare AC and MACP adsorbents at 303
AC

K are provided in Figures 5(a) and 5(b). The adsorption experiments were carried out

at pH = 5 to evaluate the removal efficiency and to avoid the formation of a precipitate

(e.g., Cu(OH)2 at pH = 6) [15,26], which is also reported for adsorption of other metal

ions [27]. Based on our preliminary studies, the effect of pH value on the adsorption

capacity of Cu2+ ions was evaluated within the pH range from 1.8 to 5.5. We observed

13
 ACCEPTED MANUSCRIPT

that the adsorption capacity of Cu2+ ions can be significantly improved with increasing

the pH value. When the adsorption was operated at low pH value, the competition

between high concentration of H+ and Cu2+ ions for the adsorptive sites on oxidized AC

intensifies, inducing weak adsorption [15]. An increase in the pH value tends to reduce

PT
H+ amount in the aqueous solution that imparts more active sites for adsorbing Cu2+

RI
ions, thus, facilitating the adsorption capacity of copper ions. The adsorption capacity

SC
of Cu ions (i.e., qe) demonstrates an increasing trend with the equilibrium concentration

(Ce) for all the adsorbents. Such a trend can be modeled using well-established
NU
adsorption models including Langmuir, Freundlich, and D-R equations.
MA

The Langmuir equation is a two-parameter model and its parameters are

tabulated in Table 2. The Langmuir parameters, qm (the amount of Cu2+ ions adsorbed
E D

per unit mass of carbon corresponding to complete monolayer coverage) and KL (the
PT

Langmuir constant related to free energy of adsorption [28,29]), were determined form
CE

the non-linearized Langmuir model [30], as shown in the remark of Table 2. These
AC

Langmuir parameters, along with the correlation coefficients (r2), of the Langmuir

equation are presented in Table 2. The trend of qm values (AF1 (23.9 mg g-1) > AF2

(20.7 mg g-1) > AC (20.2 mg g-1) > AF3 (15.5 mg g-1)) confirms the strong influence

of Fe3O4 nanoparticles loading on the removal efficiency of Cu ions from aqueous

solution. Based on the qm values, AF1 sample possesses the highest surface coverage

14
 ACCEPTED MANUSCRIPT

accessible for the liquid-phase adsorption. The KL values also demonstrate a similar

trend, indicating the presence of surface heterogeneity (due to the deposition of Fe3O4)

for the adsorption of Cu ions onto various adsorbents. As stated in the proceeding

section, both surface chemical composition and pore size distribution of MACP

PT
adsorbents have been altered, thus, resulting in different adsorption affinities toward

RI
adsorbates.

SC
The second adsorption model, Freundlich, is an empirical model that the non-

linearized from is also listed in Table 2. The KF and n values represent the Freundlich
NU
constants related to adsorption capacity and adsorption intensity, respectively. The KF
MA

and n values can be deduced from the non-linearized from of Freundlich model [30].
D

The Freundlich parameters with corresponding correlation coefficients are provided in

E

Table 2. According to Table 2, the Freundlich equation, thanks to high correlation

PT

coefficients, provides a nice fit to the experimental adsorption data. Herein the n value
CE

serves as an index determining the favorability of the adsorption process. As tabulated

AC

in Table 2, the magnitude of 1/n for all the samples is smaller than unity, predicting a

favorable adsorption of Cu ion by MACP adsorbents.

The third adsorption model, D-R equation, establishes a Gaussian function of the

Polanyi potential () between the amount adsorbed to the equilibrium concentration [29]

(the linear form is shown in Table 2). Herein the qDR (theoretical saturation capacity)

15
 ACCEPTED MANUSCRIPT

and KDR (D-R constant) values are related to the adsorption energy (E) [31]. The

parameters for the D-R equation as well as the correlation coefficients are also listed in

Table 2. According to Table 2, the AF1 sample has the most efficient adsorption

performance (based on qDR and E values) among all the samples where the adsorption

PT
capacity of Cu ions on AF1 adsorbent reaches as high as 41.6 mg g-1. It should be noted

RI
that the specific surface area is not proportional to the adsorption capacity (including

SC
qm and qDR). Therefore, the MACP adsorbent exhibits surface heterogeneity, in which

the distributions of adsorption energy and active site are mainly dominated by the
NU
oxidized carbon surface, magnetic Fe3O4, and pore size (i.e., overlapping effect). It is
MA

well-established that the electrostatic attraction is the major contributor to the

D

adsorption of metal ions, while the physical interactions and van der Waals interactions
E

play a minor role affecting the adsorption performance [5]. The magnitude of E value
PT

within the range of 7.07 to 7.99 kJ mol-1 reflects that the adsorption of Cu ions mainly
CE

originates from a chemical ion-exchange reaction rather than physical adsorption [32].
AC

Accordingly, the dispersion of magnetite nanoparticles onto AC matrix is a crucial

factor influencing the efficiency. Since the pore blockage is less significant in this case,

AF1 adsorbent not only maintains the original pore infrastructure to a great degree, but

also offers an extra magnetic surface coverage; thus providing extra active sites for the

chemical ion-exchange. In contrast, an excessive Fe3O4 loading (e.g. AF3 adsorbent)

16
 ACCEPTED MANUSCRIPT

tends to cover the main pore entrance of AC adsorbent, hindering the surface

accessibility to Cu2+ ions. Therefore, a moderate density of Fe3O4 onto AC substrate is

vital for the design of MACP adsorbents.

PT
3.3 Regeneration and recycling of MACP adsorbents

RI
The renewability of adsorbents is a principal issue since the approach of

SC
periodically regenerating adsorbents is strongly desirable in practical applications [33].

To examine the regeneration efficiency, the adsorption mechanism of Cu ions on MACP

NU
adsorbents should be clarified first. According to the analysis of D-R model, chemical
MA

ion-exchange interaction is the major contributor to the overall adsorption process. The

adsorption reactions of Cu2+ ions onto the oxidized AC adsorbent can be formulated in
E D

the presence of protons [5,34].

PT

C‒OH2+ + Cu2+ → [C‒OH Cu2+]2+ + H+ (1)

CE

C‒O‒ + Cu(OH)2‒n → [C‒O‒Cu(OH)2‒n]1‒n (2)

AC

According to Eq. (1) and (2), various oxygen functionalities (e.g., carboxyl and epoxy

groups) on the oxidized AC surface are specifically occupied by Cu ions through an

electrostatic interaction. Comparing to the MACP adsorbent, the magnetic Fe3O4

nanoparticles also provides numerous active sites for the adsorption of Cu ions.

However, the affinity towards Cu ions on oxygen functionalities is stronger than the

17
 ACCEPTED MANUSCRIPT

one on iron oxide. This difference between surface oxygen groups and magnetite is

attributed to a lower electro-negativity being offered by Fe (~1.83) compared to C

atoms (~2.55), implying that Fe-based magnetite facilitates the stripping of Cu-

adsorbed intermediates from the MACP adsorbents. Figure 6(a) shows XPS survey

PT
spectra of spent MACP adsorbent, confirming the presence of Cu species onto spent

RI
AF1 adsorbent. As illustrated in Figure 6(b), the deconvolution of Cu 2p peak on spent

SC
AF1 adsorbent composed of three main species: Cu0 at 932.8 eV (metallic Cu), Cu2+ at

934.4 eV (CuO), and Cu2+ at 935.4 eV (Cu(OH)2) [35]. Therefore, there exists an
NU
adsorption reaction sequence, which could be in progress simultaneously on oxidized
MA

AC surface (i.e., R1 and R2) and hydrophilic magnetite surface in the MACP adsorbents.
D

To measure the renewability, the adsorption/desorption cycle of Cu ions onto bare

E

AC and MACP adsorbents was repeated seven times. The variation of adsorption
PT

capacity with the cycle number is presented in Figure 7. The recovery ratio after seven
CE

adsorption/desorption cycles among various samples demonstrates an interesting trend.

AC

The bare AC sample exhibits a very poor regeneration recovery (~10%), whereas the

MACP adsorbents display an enhanced recovery efficiency (~63‒66%). Therefore, it

can be deduced that ~90% contaminates (i.e., Cu2+ ions) are still trapped on the AC

pore structure, imparting very low surface coverage for the adsorption of copper ions.

In contrast, the presence of magnetic Fe3O4 nanoparticles induces an efficient stripping

18
 ACCEPTED MANUSCRIPT

of adsorbed Cu2+ ions, regenerating the active sites of MACP surface. This can also be

inferred from the interaction between Cu2+ adsorbed complex and Fe3O4, which is

weaker compared to the electrostatic attraction between Cu2+ and oxidized AC surface.

Thus, the decoration of Fe3O4 surface onto AC matrix leads to a superior regeneration

PT
efficiency for the subsequent adsorption/desorption cycles. However, both adsorption

RI
capacity and regeneration efficiency of MACP are somehow lower than magnetic

SC
graphene derivatives, e.g., Fe3O4/carbon nanotubes: 9.36 mg/g (adsorption capacity)

and 87.5 % (regeneration efficiency after 4 cycles) [21] and Fe3O4/graphene: 32.5 mg/g
NU
(adsorption capacity) and 85% (regeneration efficiency after 8 cycles) [22]. However,
MA

granular AC is very commonly used in adsorption units (adsorption columns or towers)

D

and its price (ca. 1.5‒3.0 USD/kg) is much cheaper than graphene sheets and carbon
E

nanotubes. Thus, the MACP adsorbent still exhibits a great potential for industrial
PT

applications, based on an economic viewpoint.

CE
AC

4. Conclusions

In this work, novel MACP adsorbents were demonstrated using a facile co-

precipitation method, and the adsorption process was systematically characterized to

assess the efficiency of Cu ions removal from the aqueous solutions. With the aid of

19
 ACCEPTED MANUSCRIPT

magnetite nanoparticles, the resulting MACP adsorbent demonstrated an enhanced

adsorption capacity compared to the bare AC samples. However, an optimal density of

Fe3O4 magnetite onto AC matrix must be engineered for an efficient Cu ions removal.

Three adsorption models including Langmuir, Freundlich, and D-R used to further

PT
validate the adsorption isotherms of Cu ions on various adsorbents. The presence of

RI
magnetic Fe3O4 nanoparticles facilitated the regeneration efficiency of Cu-adsorbed

SC
intermediates from the MACP adsorbents due to a decreased electro-negativity (Fe

(~1.83) versus C atoms (~2.55)). The Fe3O4 nanoparticles also possess a strong
NU
magnetic property further enhancing the separation efficiency and regeneration of
MA

MACP adsorbents. Therefore, the MACP composites are promising for the removal of
D

heavy metals from liquid solutions, enabling a vast array of applications (e.g., drinking
E

water and wastewater).

PT
CE
AC

Acknowledgements

Financial support from the Ministry of Science and Technology of Taiwan

(MOST 105-2628-E-155-002-MY3 and MOST 105-2221-E-155-014-MY3) is

acknowledged. Partial financial support for this work under the grant number

CMRPD2E0082 by the Chang Gung Memorial Hospital, Linkou, Taiwan (Chang Gung

20
 ACCEPTED MANUSCRIPT

Medical Foundation, Taiwan) is deeply appreciated.

PT
RI
SC
NU
MA
E D
PT
CE
AC

21
 ACCEPTED MANUSCRIPT

References

[1] S. Muhammad, M.T. Shah, S. Khan, Health risk assessment of heavy metals and

their source apportionment in drinking water of Kohistan region, northern

Pakistan, Microchem. J. 98 (2011) 334-343.

PT
[2] X. Luo, X. Lei, N. Cai, X. Xie, Y. Xue, F. Yu, Removal of heavy metal ions from

RI
water by magnetic cellulose-based beads with embedded chemically modified

SC
magnetite nanoparticles and activated carbon, ACS Sustainable Chem. Eng. 4

(2016) 3960-3969.
NU
[3] Ihsanullah, A. Abbas, A.M. Al-Amer, T. Laoui, M.J. Al-Marri, M.S. Nasser, M.
MA

Khraisheh, M.A. Atieh, Heavy metal removal from aqueous solution by advanced
D

carbon nanotubes: critical review of adsorption applications, Sep. Purif. Technol.

E

157 (2016) 141-161.

PT

[4] N.K. Srivastava, C.B. Majumder, Novel biofiltration methods for the treatment
CE

of heavy metals from industrial wastewater, J. Hazard. Mater. 151 (2008) 1-8.
AC

[5] F. Fu, Q. Wang, Removal of heavy metal ions from wastewaters: a review, J.

Environ. Manage. 92 (2011) 407-418.

[6] M.R. Awual, M. Ismael, T. Yaita, S.A. El-Safty, H. Shiwaku, Y. Okamoto, S.

Suzuki, Trace copper (II) ions detection and removal from water using novel

ligand modified composite adsorbent, Chem. Eng. J. 222 (2013) 67-76.

22
 ACCEPTED MANUSCRIPT

[7] C. Liu, R. Bai, Q. San Ly, Selective removal of copper and lead ions by

diethylenetriamine-functionalized adsorbent: Behaviors and mechanisms, Water

Res. 42 (2008) 1511-1522.

[8] S. Babel, T.A. Kurniawan, Low-cost adsorbents for heavy metals uptake from

PT
contaminated water: a review, J. Hazard. Mater. 97 (2003) 219-243.

RI
[9] J. Perić, M. Trgo, N. Vukojević Medvidović, Removal of zinc, copper and lead

SC
by natural zeolite—a comparison of adsorption isotherms, Water Res. 38 (2004)

1893-1899.
NU
[10] M. Madhava Rao, A. Ramesh, G. Purna Chandra Rao, K. Seshaiah, Removal of
MA

copper and cadmium from the aqueous solutions by activated carbon derived
D

from Ceiba pentandra hulls, J. Hazard. Mater. 129 (2006) 123-129.

E

[11] Suhas, V.K. Gupta, P.J.M. Carrott, R. Singh, M. Chaudhary, S. Kushwaha,

PT

Cellulose: A review as natural, modified and activated carbon adsorbent,

CE

Bioresour. Technol. 216 (2016) 1066-1076.

AC

[12] J. Wang, H. Liu, H. Yang, C. Qiao, Q. Li, Competition Adsorption, Equilibrium,

Kinetic, and Thermodynamic Studied over La (III)-loaded Active Carbons for

Dibenzothiophene Removal, J. Chem. Eng. Data 61 (2016) 3533-3541.

23
 ACCEPTED MANUSCRIPT

[13] L. Ai, M. Li, L. Li, Adsorption of methylene blue from aqueous solution with

activated carbon/cobalt ferrite/alginate composite beads: kinetics, isotherms, and

thermodynamics, J. Chem. Eng. Data 56 (2011) 3475-3483.

[14] V.K. Gupta, S. Agarwal, A.K. Bharti, H. Sadegh, Adsorption mechanism of

PT
functionalized multi-walled carbon nanotubes for advanced Cu (II) removal, J.

RI
Mol. Liq. 230 (2017) 667-673.

SC
[15] P. Tan, J. Wen, Y. Hu, X. Tan, Adsorption of Cu2+ and Cd2+ from aqueous

solution by novel electrospun poly(vinyl alcohol)/graphene oxide nanofibers,

NU
RSC Adv. 6 (2016) 79641-79650.
MA

[16] K.V.M.K. Kireeti, C. G, M.M. Kadam, N. Jha, A sodium modified reduced

D

graphene oxide–Fe3O4 nanocomposite for efficient lead (II) adsorption, RSC

E

Adv. 6 (2016) 84825-84836.

PT

[17] K.C. Justi, V.T. Fávere, M.C.M. Laranjeira, A. Neves, R.A. Peralta, Kinetics and
CE

equilibrium adsorption of Cu(II), Cd(II), and Ni(II) ions by chitosan

AC

functionalized with 2[-bis-(pyridylmethyl)aminomethyl]-4-methyl- 6-

formylphenol, J. Colloid Interface Sci. 291 (2005) 369-374.

[18] K. Kadirvelu, C. Faur-Brasquet, P.L. Cloirec, Removal of Cu (II), Pb (II), and Ni

(II) by adsorption onto activated carbon cloths, Langmuir 16 (2000) 8404-8409.

[19] B. Biškup, B. Subotić, Removal of heavy metal ions from solutions using zeolites.

24
 ACCEPTED MANUSCRIPT

III. influence of sodium ion concentration in the liquid phase on the kinetics of

exchange processes between cadmium ions from solution and sodium ions from

zeolite A, Sep. Sci. Technol. 39 (2005) 925-940.

[20] A.L. Cazetta, O. Pezoti, K.C. Bedin, T.L. Silva, A. Paesano Junior, T. Asefa, V.C.

PT
Almeida, Magnetic activated carbon derived from biomass waste by concurrent

RI
synthesis: efficient adsorbent for toxic dyes, ACS Sustainable Chem. Eng. 4

SC
(2016) 1058-1068.

[21] Z.-F. Yang, L.-Y. Li, C.-T. Hsieh, R.-S. Juang, Co-precipitation of magnetic Fe3O4
NU
nanoparticles onto carbon nanotubes for removal of copper ions from aqueous
MA

solution. J. Taiwan Inst. Chem. Engineers 82 (2018) 56-63.

D

[22] Z.-F. Yang, L.-Y. Li, C.-T. Hsieh, R.-S. Juang, Y. Ashraf Gandomi, Fabrication of
E

magnetic Iron oxide@graphene composites for adsorption of copper ions from

PT

aqueous solutions, Mater. Chem. Phys. 219 (2018) 30-39.

CE

[23] W. Fan, W. Gao, C. Zhang, W.W. Tjiu, J. Pan, T. Liu, Hybridization of graphene
AC

sheets and carbon-coated Fe3O4 nanoparticles as a synergistic adsorbent of

organic dyes, J. Mater. Chem. 22 (2012) 25108-25115.

[24] W.H. Lee, J.Y. Kim, Y.K. Ko, P.J. Reucroft, J.W. Zondlo, Surface analysis of

carbon black waste materials from tire residues, Appl. Surf. Sci. 141 (1999) 107-

113.

25
 ACCEPTED MANUSCRIPT

[25] T. Takahagi, A. Ishitani, XPS studies by use of the digital difference spectrum

technique of functional groups on the surface of carbon fiber, Carbon 22 (1984)

43-46.

[26] A.B. Dichiara, M.R. Webber, W.R. Gorman, R.E. Rogers, Removal of copper

PT
ions from aqueous solutions via adsorption on carbon nanocomposites, ACS

RI
Appl. Mater. Interfaces 7 (2015) 15674-15680.

SC
[27] S. Handley-Sidhu, T.K. Mullan, Q. Grail, M. Albadarneh, T. Ohnuki, L.E.

Macaskie, Influence of pH, competing ions, and salinity on the sorption of

NU
strontium and cobalt onto biogenic hydroxyapatite, Sci. Reports 6 (2016) 23361.
MA

[28] H. Teng, C.T. Hsieh, Influence of surface characteristics on liquid-phase

D

adsorption of phenol by activated carbons prepared from bituminous coal, Ind.

E

Eng. Chem. Res. 37 (1998) 3618-3624.

PT

[29] C.T. Hsieh, H. Teng, Influence of mesopore volume and adsorbate size on
CE

adsorption capacities of activated carbons in aqueous solutions, Carbon 38 (2000)

AC

863-869.

[30] K.V. Kumara, K. Porkodi, F. Rocha, Comparison of various error functions in

predicting the optimum isotherm by linear and non-linear regression analysis for

the sorption of basic red 9 by activated carbon, J. Hazard. Mater. 150 (2008)

158-165.

26
 ACCEPTED MANUSCRIPT

[31] R.C. Reid, J.M. Prausnitz, B.E. Poling, The properties of gases and liquids,

McGraw-Hill Book Company, New York, 1987.

[32] R. Chen, L. Chai, Q. Li, Y. Shi, Y. Wang, A. Mohammad, Preparation and

characterization of magnetic Fe3O4/CNT nanoparticles by RPO method to

PT
enhance the efficient removal of Cr (VI), Environ. Sci. Pollut. Res. 20 (2013)

RI
7175-7185.

SC
[33] S. Li, Y. Gong, Y. Yang, C. He, L. Hu, L. Zhu, L. Sun, D. Shu, Recyclable

CNTs/Fe3O4 magnetic nanocomposites as adsorbents to remove bisphenol A

NU
from water and their regeneration, Chem. Eng. J. 260 (2015) 231-239.
MA

[34] X. Ren, C. Chen, M. Nagatsu, X. Wang, Carbon nanotubes as adsorbents in

D

environmental pollution management: a review, Chem. Eng. J. 170 (2011) 395-

E

410.
PT

[35] C.D. Wagner, Handbook of X-ray photoelectron spectroscopy: a reference book

CE

of standard data for use in X-ray photoelectron spectroscopy, Physical Perkin-

AC

Elmer Corp/Electronics Division, New York, (1979), pp. 86-87.

27
 ACCEPTED MANUSCRIPT

Table 1. Surface characteristics of bare AC and MACP adsorbents, determined by

nitrogen physisorption at -196°C.

Textural
AC AF1 AF2 AF3
properties
(SBET)-Surface
647 516 444 202
area (m2 g-1)

PT
(VT)-total pore
0.452 0.452 0.392 0.386
volume (cm3 g-1)
(Vμ)-micropore

RI
0.152 0.113 0.096 0.020
volume (cm3 g-1)
(Vm)-mesopore

SC
0.264 0.321 0.280 0.366
volunme (cm3 g-1)
Percentage of
33.6 25.0 24.4 5.1
NU
micropore (%)
Percentage of
58.4 75.0 75.6 94.9
mesopore (%)
MA

Average pore size

28.0 35.0 35.3 76.4
(nm)
E D
PT
CE
AC

28
 ACCEPTED MANUSCRIPT

Table 2. Langmuir, Freundlich, and D-R parameters, together with correlation

coefficients, for adsorption of Cu ions on bare AC and MACP adsorbents at

30°C.

Langmuir model Freundlich model D-R model

Sample
ID qm KL KF qDR E

PT
R2 n R2 R2
(mg g-1) (L mg-1) (mg1-(1/n)L3/ng-1) (mg g-1) (kJ mol-1)

AF1 23.9 0.003 0.971 1.94 0.520 0.987 41.6 7.99 0.982

RI
AF2 20.7 0.002 0.982 1.63 0.191 0.985 35.7 7.26 0.985

SC
AF3 15.5 0.001 0.976 1.55 0.121 0.976 27.1 7.07 0.975

AC 20.2 0.002 0.981 1.96 0.455 0.990 35.5 7.98 0.991

NU
Remark:
K L qmCe
1. Langmuir model: qe  , where qL is the maximal adsorption capacity
1  K LCe
MA

and KL is the Langmuir constant.

2. Freundlich model: qe  K FCe , where KF and n are the Freundlich constants,

1/ n
D

respectively.
E

1
3. D-R model: qe  qDR exp[  B 2 ] with   RT ln(1  ) , where qDR is the
PT

Ce
theoretical saturation capacity, R is the gas constant, T is the absolute temperature,
CE

and KDR is the D-R constant, related to the adsorption energy.

AC

29
 ACCEPTED MANUSCRIPT

Figures Caption

Figure 1. Schematic diagram illustrating (a) synthesis of MACP adsorbent and (b)

removal of Cu ions from aqueous solution by the MACP adsorbent.

Figure 2. FE-SEM images of MACP adsorbent with (a) low- and (b) high-

PT
magnification. (c) Typical XRD patterns of different adsorbents.

RI
Figure 3. (a) Survey-scan XPS spectra of different adsorbents. (b) O 1s, (c) Fe 2p,

SC
and (d) C 1s of MACP adsorbent, deconvoluted by multiple Gaussian

function.
NU
Figure 4. TGA curves of MACP adsorbent with the entire temperature range between
MA

30 and 800°C.
D

Figure 5. Equilibrium adsorption isotherms of Cu ions on bare AC and MACP

E

adsorbents, fitted with (a) Freundlich and Langmuir models and (b) D-R
PT

model.
CE

Figure 6. (a) Survey-scan and (b) Cu 2p peak of spent MACP adsorbents.

AC

Figure 7. Variation of equilibrium adsorption capacity with recycle number for bare

AC and MACP adsorbents.

30
 ACCEPTED MANUSCRIPT

(a) (b)
Adsorption
Fe3+: Fe2+ = 2:1
4M NH3·H2O
HNO3 (pH > 10) T= 303 K, 2 h Magnetic
separation
MACP

PT
Desorption
Magnetic

RI
Activated carbon particle
activated
carbon Recycling
particle
(MACP)

SC
Analysis

Removal of MACP

Activated carbon
NU
(Irregularly arranged graphite)
MA

Figure 1. Schematic diagram illustrating (a) synthesis of MACP adsorbent and (b)

removal of Cu ions from aqueous solution by the MACP adsorbent.

E D
PT
CE
AC

31
 ACCEPTED MANUSCRIPT

(a) (c)

Oxidized AC

PT
(b)

RI
SC
NU
MA

Figure 2. FE-SEM images of MACP adsorbent with (a) low- and (b) high-

magnification. (c) Typical XRD patterns of different adsorbents.

E D
PT
CE
AC

32
 ACCEPTED MANUSCRIPT

(a) (c)

Oxidized AC

PT
RI
SC
(b) NU (d)
MA
E D
PT

Figure 3. (a) Survey-scan XPS spectra of different adsorbents. (b) O 1s, (c) Fe 2p,

and (d) C 1s of MACP adsorbent, deconvoluted by multiple Gaussian

CE

function.
AC

33
 ACCEPTED MANUSCRIPT

PT
AF3

RI
AF2

SC AF1
NU
MA
E D

Figure 4. TGA curves of MACP adsorbent with the entire temperature range
PT

between 30 and 800°C.

CE
AC

34
 ACCEPTED MANUSCRIPT

(a)

PT
RI
SC
NU
(b)
MA
E D
PT
CE
AC

Figure 5. Equilibrium adsorption isotherms of Cu ions on bare AC and MACP

adsorbents, fitted with (a) Freundlich and Langmuir models and (b) D-R

model.

35
 ACCEPTED MANUSCRIPT

(a)

PT
RI
SC
NU
(b)
MA

Metallic Cu Cu(OH)2
E D

CuO
PT
CE
AC

Figure 6. (a) Survey-scan and (b) Cu 2p peak of spent MACP adsorbents.

36
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
E D
PT

Figure 7. Variation of equilibrium adsorption capacity with recycle number for bare

AC and MACP adsorbents.

CE
AC

37
 ACCEPTED MANUSCRIPT

Graphical Abstract

Adsorption
Fe3+: Fe2+ = 2:1
4M NH3·H2O
HNO3 (pH > 10) T= 303 K, 2 h Magnetic
Synthesis of separation
MACP adsorbent MACP Adsorption/desorption
of Cu2+ ions

PT
Desorption
Activated carbon particle Magnetic
activated
carbon Recycling

RI
particle
(MACP)
Analysis

Removal of MACP

SC
Activated carbon
(Irregularly arranged graphite)
NU
MA
E D
PT
CE
AC

38
 ACCEPTED MANUSCRIPT

Research Highlights

 Magnetic activated carbon particle is fabricated by a co-precipitation method.

 The particles serve as an adsorbent for removal of Cu ions from aqueous

solution.

PT
 A moderate density of magnetite onto carbon is crucial for removing copper ions.

RI
 The magnetic particles facilitate adsorption and regeneration efficiencies of ions.

SC
 The design of adsorbent delivers a potential for environmental applications.
NU
MA
E D
PT
CE
AC

39