Fabrication of a GNP/Fe-Mg Binary Oxide Composite for Effective Removal of Arsenic from Aqueous Solution.

Graphene nanoplates (GNPs) can be used as a platform for homogeneous distribution of adsorbent nanoparticles to improve electron exchange and ion transport for heavy-metal adsorption. In this study, we report a facile thermal decomposition route to fabricate a graphene-supported Fe-Mg oxide composite. The prepared composite was characterized using scanning electron microscopy, transmission electron microscopy, energy-dispersive spectrometry, X-ray diffraction, and X-ray photoelectron spectroscopy. Batch experiments were carried out to evaluate the arsenic adsorption behavior of the GNP/Fe-Mg oxide composite. Both the Langmuir and Freundlich models were employed to describe the adsorption isotherm, in which the sorption kinetics of the arsenic adsorption process by the composite was found to be pseudo-second-order. Furthermore, the reusability and regeneration of the adsorbent were investigated by an assembled-column filter test. The GNP/Fe-Mg oxide composite exhibited significant fast adsorption of arsenic over a wide range of solution pHs, with exceptional durability and recyclability, which could make this composite a very promising candidate for effective removal of arsenic from aqueous solutions.

Introduction

Contamination of water with toxic metals, such as arsenic (As), lead (Pb), and mercury (Hg), has significantly increased over the last few decades due to anthropogenic sources (industrialization and urbanization). Because of their toxicity and carcinogenicity to human beings, exposure to these elements poses a serious health risk.[1] Among them, arsenic is one of the most toxic elements and is widely present in nature through leaching from soils, mining, fertilizers, industrial waste, biological activity, and As-containing minerals.[2,3] As a consequence, the World Health Organization (WHO) has established a maximum recommended concentration of arsenic in water of 10 ppb. Long-term ingestion of food grown in arsenic-contaminated areas or directly drinking arsenic-contaminated water is linked to kidney, skin, and lung cancers.[4] Therefore, it is urgent to remove arsenic from contaminated water to provide safe drinking water with arsenic levels below the limit recommended by the WHO. There are several approaches that have been used to remove arsenic from water, including adsorption, ion exchange, reverse osmosis, electrochemical treatment, membrane filtration, and co-precipitation.[5−8] However, because of its simplicity, low cost, and high efficiency, adsorption is widely employed and studied as a promising technology for cost-effective arsenic removal.

Many materials have been employed as adsorbents for arsenic adsorption, such as agricultural and industrial wastes, surfactants, carbon-based materials, polymers, and metal oxides.[9,10] Among these, metals and metal oxides, such as TiO2,[11−13] nano zero-valent iron,[14,15] Fe2O3,[4,16,17] Fe3O4,[18] CeO2,[19] CuO,[20,21] CaO,[22] and ZrO2,[23,24] have been extensively studied for arsenic treatment in aqueous solutions because of their high affinity to the arsenic species, low cost, and tunability of adsorption capacity.[10,25] Recently, considerable attention has been focused on the development of adsorbent composites containing two or more metals and metal oxides to maximize arsenic adsorption. For example, Shan and Tong fabricated Fe–Mn binary oxides with a high adsorption capacity for arsenic.[26] Also, superparamagenetic Mg0.27Fe2.5O4, a novel arsenic adsorbent, was synthesized by Tang et al.[27] In another report, Yu et al. presented Fe–Ti binary oxide magnetic nanoparticles, which combined the photocatalytic oxidation property of TiO2 with the high adsorption capacity and magnetic properties of γ-Fe2O3 for arsenic treatment.[28] Xu et al. reported the synthesis and application of a CeO2–ZrO2 composite for the removal of arsenic from aqueous solutions.[29] Basu and Ghosh found that Fe(III)–Al(III) mixed oxides and Fe(III)–Ce(IV) oxides have a high adsorption capacity toward arsenic.[30,31]

Graphene, a two-dimensional material, has been attracting significant interest over the past decade due to its exceptional chemical and physical properties, which can be applied to many different areas, including, but not limited to, electronic devices, energy storage and conversion, sensors, adsorption, and composites.[32−37] Most recently, graphene has gained tremendous interest as a supporting material for enhancement of the adsorption properties of adsorbents due to its large surface area, high conductivity, ionic mobility, and superior mechanical flexibility. For example, Gollavelli et al. reported a smart magnetic graphene that removed heavy metals from drinking water.[38] A hybrid of monolithic Fe2O3/graphene was also fabricated and showed favorable properties for arsenic removal.[39] Reduced graphene oxide-supported mesoporous Fe2O3/TiO2 nanoparticles synthesized by a sol–gel route showed high adsorption toward arsenic.[40] Kumar et al. synthesized single-layer graphene oxide with manganese ferrite magnetic nanoparticles for efficient removal of arsenic from contaminated water.[41]

Herein, we report a simple one-pot hydrothermal method to prepare a GNP-supported Fe–Mg binary oxide (GNP/Fe–Mg oxide) composite. The obtained material was well characterized by transmission electron microscopy (TEM), scanning electron microscopy (SEM), energy-dispersive spectrometry (EDS), thermogravimetric analysis (TGA), X-ray photoelectron spectroscopy (XPS), and X-ray diffraction (XRD). Arsenic adsorption was employed to evaluate the adsorption capacity of the adsorbent. The effects of the parameters, including the Fe/Cu ratio, graphene loading, initial arsenic concentration, adsorption time, and solution pH, on arsenic adsorption were investigated through batch experiments, and a column test was also used to study the recyclability of the sorbent.

Results and Discussion

The morphologies of the prepared graphene nanoplates (GNPs) and free-standing Fe–Mg oxide were studied by SEM, and the images are shown in Figure . It can be clearly seen in Figure A that the obtained GNPs have a crumpled, wrinkled morphology, with a diameter of tens of microns and thickness <20 nm.[42] The SEM image of the as-prepared Fe–Mg binary oxide is shown in Figure B, which indicates that the Fe–Mg oxide consists of ultrafine nanocrystallites. The morphology of the Fe–Mg oxide is consistent with that in a previous study.[27] We believe that the metal ions first physically adsorbed onto the GNPs and then reacted under hydrothermal conditions to form metal oxides on the GNPs.

Figure 1

SEM images of (A) GNPs and (B) free-standing Fe–Mg oxide.

SEM and TEM images of the GNP/Fe–Mg oxide composite are shown in Figure . In the low-magnification SEM and TEM images (Figure A–C), it is clear that Fe–Mg oxides are well formed and dispersed on the surface of GNPs. When the composite is viewed at higher SEM and TEM magnifications, as shown in Figure D and insets of Figure B,D, it can be seen that Fe–Mg oxides on graphene are not well-defined nanocrystallites, with some agglomeration observed, which might be ascribed to the amorphous nature of Fe–Mg binary oxides on the GNPs. The amorphous nature of the Fe–Mg oxide formed on the GNPs is further confirmed by the XRD pattern of the composite (Figure S1). In the XRD pattern, apart from the peaks with an asterisk, which are attributed to the crystallites of GNPs,[42,43] the presence of very weak and broad peaks in region of 30–80° indicates the amorphous nature of the Fe–Mg oxide.

Figure 2

(A, B) SEM images and (C, D) TEM images of the GNP/Fe–Mg binary oxide composite.

EDS mapping was undertaken to study the elemental distribution in the GNP-supported Fe–Mg oxide composite, as exhibited in Figure . It can be clearly seen from the figure that the Mg–Fe oxide uniformly covers the entire GNP surface.

Figure 3

EDS mapping of the GNP/Fe–Mg binary oxide composite: (A) overlapping elements; (B) C; (C) Mg; and (D) Fe.

To explore the chemical states of C, Fe, Mg, and O in the composite, the core-level XPS spectra of C 1s, Fe 2p, Mg 1s, and O 1s and their corresponding deconvolutions showing the contributions of the individual components are shown in Figure . The C1 s spectrum (Figure A) has two major peaks at 284.1 and 284.8 eV, which correspond to carbon in C=C (sp2) and C–C (sp3) bonds,[44] respectively, of graphenic carbons. Deconvoluted core-level Fe 2p shows two dominant peaks at 711.4 and 724.5 eV, with satellites corresponding to Fe 2p3/2 and Fe 2p1/2 of the Fe3+ state, respectively.[45]Figure C shows the binding energy and its deconvolution spectrum for core-level Mg 1s. There is only one peak observed at 1303.5 eV, which could be attributed to MgO.[46] In the O 1s spectrum (Figure D), the peaks at 530.1 and 531.4 eV are due to the presence of O2– in Fe2O3 and MgO, respectively.[39,47] Interestingly, the appearance of the peak at 532.2 eV demonstrates the existence of Fe–OH in the composite, which enhances the adsorption capacity for heavy ions.[48,49] There is also the existence of a peak at 533.8 eV, which is ascribed to the Si–O bond used as a substrate for the XPS investigation.

Figure 4

Core-level XPS spectra of C 1s (A), Fe 2p (B), Mg 1s (C), and O 1s (D) obtained for the GNP/Fe–Mg binary oxide composite.

The thermal properties of the prepared GNP-supported Fe–Mg oxide composite were investigated by TGA, as presented in Figure . The TGA curve shows a two-step decomposition process. The first weight loss in the range of 30–150 °C is 0.52%, which can be attributed to the removal of adsorbed and trapped water in the composite. The weight loss of 0.93% in the second step (300–550 °C) is due to the decomposition of Fe(OH)3 to α-Fe2O3.

Figure 5

TGA of the GNP/Fe–Mg binary oxide composite.

Figure A shows the effect of various Mg/Fe weight ratios on arsenic adsorption, with an initial As(V) concentration of 10 mg/L, an adsorbent dose of 200 mg/L, a pH of 7, and at room temperature. The results indicate that As(V) sorption by GNP-supported Fe–Mg oxide significantly increases with a decrease in the Mg/Fe weight ratio, and it reaches a maximum equilibrium adsorption capacity of 40.3 mg/g when the Mg/Fe ratio is 5:5. Then, As(V) adsorption slowly declines with a further decrease in the Mg/Fe weight ratio. It is also of note that As(V) sorption by GNP–Fe2O3 (31 mg/L) is somewhat higher than that by GNP–MgO (2 mg/L). The effect of the GNP load on the As(V) sorption capacity by the composite was also investigated, with an initial As(V) concentration of 9.5 mg/L, as shown in Figure B. It is obvious that when the GNP loading increases the As(V) sorption increases, and it reaches a maximum of approximately 37.5 mg/g at the GNP/Mg–Fe oxide weight ratio of 2:1. When the GNP load is further increased, the sorption capacity significantly decreases, and without Fe–Mg oxide, the As(V) sorption capacity by pure GNPs is only 2.38 mg/g. These results suggest that the combination of GNPs and Fe–Mg oxide significantly improves As(V) adsorption.

Figure 6

Effect of (A) the Mg/Fe weight ratio and (B) GNP loading on As(V) adsorption by GNP/Fe–Mg oxide composites.

The adsorption isotherm was obtained to assess the arsenic adsorption and determine the maximum As(V) adsorption capacity of the GNP/Fe–Mg oxide composite. The amount of arsenic adsorbed onto the composite at equilibrium (qe) was calculated from the As(V) concentration difference, with the following equationwhere C0 (mg/L) is the initial concentration, Ce (mg/L) is the equilibrium concentration, V (L) is the solution volume, and m (g) is the mass of the GNP/Fe–Mg oxide adsorbent.

Figure shows the arsenic adsorption capacity of the composite at equilibrium at various As(V) concentrations, in range of 5–90 mg/L; an adsorbent dose of 200 mg/g; pH 7; and at room temperature. The Langmuir and Freundlich adsorption isotherm models were employed to fit the data, as expressed in eqs and 3, respectivelywhere qe is the amount of arsenic adsorbed onto the solid phase at equilibrium (mg/g), qmax (mg/g) is the maximum arsenic adsorption capacity per unit weight of adsorbent, Ce is the equilibrium arsenic concentration (mg/L), KL is the equilibrium adsorption constant representing the affinity of binding sites (L/mg), KF is the Freundlich constant, and n is the heterogeneity factor.

Figure 7

Adsorption isotherm for As(V) adsorption by the GNP/Fe–Mg binary oxide composite.

The obtained adsorption constants are presented in Table . The higher correlation coefficient (0.985) value of the fitted Freundlich plot compared to that of the Langmuir plot (0.935) suggests that the Freundlich model is more suitable for representing the adsorption behavior of As(V) by the GNP/Fe–Mg oxide composite than the Langmuir model. The low calculated heterogeneity factor (n = 0.4) also suggests that the Freundlich model is the more favorable model to describe arsenic adsorption by GNP/Fe–Mg oxide. These results indicate that As(V) is heterogeneously adsorbed onto the composite surface, which could be attributed to the simultaneous existence of graphene, iron, and magnesium oxides in the solid phase. The maximum As(V)-adsorption capacity of GNP/Fe–Mg oxide determined from the Langmuir model is 103.9 mg/g, which indicates that the composite is very effective for arsenic removal. Illustrated in Table is a comparison of the As(V) adsorption capacity of the GNP/Fe–Mg oxide composite with that of other adsorbents. It is obvious from the comparison that the adsorption capacity of GNP/Fe–Mg oxide outperforms that of the majority of other adsorbents, which could make this material a promising sorbent for arsenic removal.

Table 1

Langmuir and Freundlich Isotherm Parameters for As(V) Adsorption on the GNP/Fe–Mg Binary Oxide Composite

Langmuir model			Freundlich model
qmax (mg/g)	KL (L/mg)	R2	KF	n	R2
103.9	0.05	0.935	13.57	0.4	0.985

Table 2

Comparison of Maximal Arsenic Adsorption Capacity by Various Adsorbents

adsorbates	pH	qmax (mg/g)	references
superparamagnetic Mg0.27Fe2.5O4	7	83.2	(27)
Fe3O4–GO (MGO)	6.5	59.6	(50)
FeMnOx/RGO	7	22.22	(51)
CeO2–grahene composite	4	1.019	(28)
GO–ZrO(OH)2 nanocomposites	5–11	84.89	(52)
nZVI/graphene	7	29	(53)
magnetic graphene	4	3.26	(38)
Fe3O4/graphene/LDH	6	73.1	(54)
magnetic-GO	4	38	(55)
magnetic-rGO	4	12	(55)
MnFe2O4	3	90	(58)
CoFe2O4	3	74	(58)
GO/MnFe2O4	4	207	(41)
GNP/Fe–Mg oxide	7	103.9	this work

The adsorption kinetics of As(V) was obtained to further understand the adsorption behavior of As(V) on the GNP/Fe–Mg oxide surface. Figure shows the adsorption data with an initial As(V) concentration of 5 mg/L at different time intervals. The adsorption process reaches equilibrium within 3 h. The pseudo-second-order model was applied to describe the kinetic data, as expressed in eq where qt (mg/g) is the amount of arsenic adsorbed on the solid phase at time t (h), qe (mg/g) is the amount of arsenic adsorbed on the solid phase at equilibrium, and K is the adsorption rate constant (g mg h). According to the adsorption kinetic values listed in Table , the experimental data are relatively well fitted, with a correlation coefficient of 0.957. This result implies that the adsorption process might be chemical adsorption accompanied by electron exchange between the composite and arsenic.[56] Furthermore, the experimental adsorption capacity at equilibrium of the composite (22.4 mg/g) with an initial As(V) concentration of 5 mg/L is close to the calculated value from the pseudo-second-order model (25.5 mg/g).

Figure 8

Adsorption kinetics of As(V) on the GNP/Fe–Mg binary oxide composite.

Table 3

Adsorption Kinetics Parameters for As(V) Adsorption on the GNP/Fe–Mg Binary Oxide Composite

pseudo-second-order model
qe (mg/g)	K (h–1)	R2
25.5	0.227	0.957

The effect of solution pH on arsenic adsorption by the GNP/Fe–Mg oxide composite was determined with an initial As(V) concentration of 10 mg/L, as shown in Figure . It can be clearly seen that As(V) sorption is significantly dependent upon the pH. Under acidic conditions, arsenic adsorption slowly increases with an increase in solution pH, and it reaches a maximum capacity under neutral conditions (pH 7). The adsorption capacity significantly decreases under mild basic conditions (pH 8–9) before greatly declining when the solution pH further increases. This phenomenon may be attributed to the dependence of the adsorption of strong acid anions by metal oxides and hydroxides on solution pH.[57] The dissolution of Fe and Mg in the GNP/Fe–Mg oxide composite at pH 7 was also recorded, as shown in Figure S2. It is obvious that the release of Fe and Mg is minimal compared with As(V) adsorption, which indicates that the GNP/Fe–Mg binary oxide composite is a stable and effective adsorbent for arsenic. As the source of GNPs in this study does not contain −COOH or −OH groups, the adsorption of arsenic on GNPs is mostly physical adsorption. However, when Mg–Fe oxide is loaded onto GNPs, there exists −OH groups on the surface of the metal oxide in aqueous solution (especially under the optimum pH condition). These −OH groups are responsible for the dramatic increase in arsenic adsorption due to electrostatic attraction.

Figure 9

Effect of solution pH on As(V) adsorption by the GNP/Fe–Mg binary oxide composite.

As electrostatic attraction is a main force, which is responsible for the adsorption of As(V) on graphene–metal oxide composites,[50,52] the change in the electrostatic force between As(V) and the GNP/Fe–Mg binary oxide composite may explain the effect of pH on As(V) adsorption. At a low pH, the adsorbent of GNP/Fe–Mg oxide has a net positive charge due to protonation of the −OH groups in the Fe–Mg binary oxide. As a result, they attract negatively charged As(V) anions, which leads to greater adsorption. When the pH increases, the positive charge decreases, resulting in a decrease in As(V) adsorption.

The selectivity of the GNP/Fe–Mg binary oxide adsorbent toward arsenic in the presence of common metal ions that exist in drinking water, such as Na+, K+, Ca2+, and Mg2+, was studied (Figure S3). It can be seen that whereas more than 95% of the arsenic was adsorbed there was an insignificant amount of Na+, K+, and Ca2+ ions adsorbed onto the GNP/Fe–Mg oxide composite. Interestingly, the concentration of Mg2+ in the tested solution increased by 10% after adsorption due to the release of Mg from the adsorbent.

To evaluate the recyclability, a filter column with a diameter of 2 cm and height of 10 cm was assembled, as shown in Figure . The mass of the GNP/Fe–Mg oxide composite used was 200 mg. The adsorption process was carried out with 10 mL of flow solution (pH 7) of 3 mg/L As(V). After adsorption, the filter column was washed with 20 mL of 2 M NaOH solution to regenerate the adsorbent before the next test cycle. The adsorption–regeneration process was repeated for five cycles. An insignificant decrease in the removal efficiency (less than 2%) after five cycles suggests that the GNP/Fe–Mg oxide composite has a high durability for arsenic removal.

Figure 10

Recyclability of the GNP/Fe–Mg binary oxide composite for As(V) removal in a column test.

Experimental Section

Materials

Natural graphite flakes were supplied from Pressol Gmbh, with particle size >100 mesh. Chemicals, such as dry acetone, ethanol, concentrated sulfuric acid (98%), sodium hydroxide (NaOH), potassium hydroxide (KOH), ethanol, sodium persulfate (Na2S2O8), MgCl2·6H2O, and anhydrous FeCl3, were purchased from Ajax Finechem. All chemicals were used as received.

Synthesis of GNPs

Preparation of GNPs was undertaken according to previous methods.[42] Typically, 1 g of natural graphite flakes was dispersed in 80 mL of sulfuric acid (98%), with swirling for 10 min, in a 250 mL reactor. Thereafter, 10 g of sodium persulfate (Na2S2O8) was gradually added and the reaction mixture was gently stirred for 3 h at room temperature. The mixture was then filtered with a sintered glass filter (Duran sintered with a disc filter funnel of capacity 50 mL; maximum pore size, 4–5 μm) without quenching. The filtrate solution was separated and used for graphene exfoliation. The GNPs thus obtained were rinsed with dry acetone (3 × 10 mL) followed by 5 mL of water to remove any residual acid on the GNPs and dried at 60 °C in open air.

Synthesis of the GNP/Fe–Mg Binary Oxide Composite

GNP/Fe–Mg binary oxide composites were fabricated by a simple one-pot hydrothermal strategy. First, MgCl2·6H2O and FeCl3 with various Mg/Fe weight ratios, 10:0, 9:1, 7:3, 5:5, 3:7, 1:9, and 0:10, were dissolved in 50 mL of ethanol. Then, GNPs with different loadings were dispersed in the solution mixture by sonication for 10 min and stirring for 1 h. Subsequently, a 2 M NaOH solution was added dropwise into the solution under vigorous stirring until a pH of ∼8 or 9 was reached. After 1 h of further stirring, the reaction solution was transferred and sealed in a Teflon-lined autoclave and placed in an oven preheated to 150 °C for 2 h. Then, the solution was cooled to room temperature and the precipitate was filtered and washed three times each with ethanol and distilled water. The sample was dried overnight at a temperature of 60 °C in air to obtain the GNP/Fe–Mg oxide composites.

Characterization

The structures and mapping elemental compositions of the samples were studied by EDS-equipped scanning electron microscopy (SEM), using an FEI Nova NanoSEM (Hillsboro) operating under high-vacuum conditions, with an accelerating voltage of 30 keV and an Everhart–Thornley detector. TEM images were obtained on a JEOL 1010 TEM instrument, operated at an accelerating voltage of 100 kV. Thermal analysis was performed on a TGA instrument from Perkin-Elmer, with a furnace, a microbalance, an operating temperature from 35 to 800 °C, and a heating rate of 20 °C/min. A BrukerAXS D8 Discover instrument with a general area detector diffraction system using a Cu Kα source was utilized to obtain XRD patterns. XPS spectra were obtained on a K-Alpha XPS instrument using monochromated aluminum as the X-ray source. C 1s, Fe 2p, Mg 1s, and O 1s core-level spectra were recorded with an overall resolution of 0.1 eV. The core-level spectra were background-corrected using the Shirley algorithm, and chemically distinct species were resolved using a nonlinear least-squares fitting procedure.

Adsorption Studies

A stock solution of 1000 ppm As(V) was prepared by dissolving As2O5 in water. The arsenic concentration was determined using an Agilent 4200 microwave plasma-atomic emission spectrometer (MP-AES). All samples were analyzed within 24 h after filtration.

Effects of the Mg/Fe Weight Ratio and GNP Loading on Arsenic Sorption

Adsorption experiments were carried out in closed glass vessels. Typically, 10 mg of each adsorbent, prepared from different Mg/Fe weight ratios and various GNP loadings, was added into glass vessels containing 50 mL of 10 mg/L arsenic solution at pH 7. The solution was kept shaking at 200 rpm at room temperature for 24 h. Then, all samples were filtered to remove the adsorbent and the concentration of arsenic in the residual solutions was analyzed.

Adsorption Isotherm

In glass vessels, 10 mg of the optimally fabricated adsorbent was added to 50 mL of arsenic solution, with initial concentrations ranging from 5 to 90 mg/L. Adsorption was carried out at a solution pH of 7 at room temperature, stirring at a speed of 200 rpm for 24 h. The mixtures were then filtered and analyzed for residual arsenic by MP-AES.

Adsorption Kinetics

In a typical experiment, 40 mg of the GNP/Fe–Mg binary oxide was mixed with 200 mL of 5 mg/L arsenic in a glass vessel. The mixed solution was shaken at 200 rpm on an orbital shaker, at room temperature and pH 7. At certain time intervals, 10 mL of the mixture was taken, filtered, and analyzed for arsenic.

Effect of Solution pH

The GNP/Fe–Mg oxide composite (10 mg) was added to 50 mL of 10 mg/L arsenic at various solution pH values, ranging from 4 to 11 (the pH values were adjusted with dilute HCl and NaOH solutions). The suspensions were shaken at a speed of 200 rpm at room temperature for 24 h. Then, all samples were filtered and the residual arsenic concentrations were determined.

Recyclability Test

The reusability of the prepared adsorbent was studied by a column test. The GNP/Fe–Mg oxide composite was assembled as a part of a filter column. Other parts included a glass tube with both ends wrapped with a few layers of filter paper and cotton. After each arsenic adsorption cycle, the filter column was regenerated by washing it several times with 2 M NaOH before implementing the next experiment.