One-pot hydrothermal synthesis of carbonaceous nanocomposites for efficient decontamination of copper

Abstract

Environmental pollution associated with heavy metal ions has always been a serious environmental problem. To date, carbonaceous adsorbents are still the most promising candidates for the decontamination of these ions, but are limited by either cost or adsorption capacity. Here, carbonaceous nanocomposites containing numerous surface acidic groups were easily synthesized via one-pot hydrothermal carbonization of glucose in the presence of acrylic acid and a small amount of graphene oxide, and characterized in detail. Batch adsorption of Cu(II) on the carbonaceous nanocomposites from aqueous solution showed that these materials exhibited an excellent adsorption affinity for Cu(II) and the maximum adsorption capacity was as high as 146.1 mg g⁻¹ at pH 5 and T = 298 K, which was much higher than any previous reports. The effect of the degree of functionality on adsorption behaviors, as well as the effects of pH, ionic strength, complex anions, temperature, and the presence of natural organic compounds (humic acid and fulvic acid) and organic pollutants (ionic liquid), were studied systematically to understand the adsorption mechanism. In addition, X-ray photoelectron spectroscopy was further used to confirm that surface complexation reactions played an important role in the adsorption process. This work would provide cheap nanocomposites which could be candidate materials for efficient decontamination of copper ions.

---

Introduction

Heavy metal ion pollution in industrial wastewater has attracted global attention due to its adverse effects on the environment and human health.^1,2 Copper, an abundant and naturally occurring element present in municipal wastewaters,³ is one of the most widespread and common heavy metal contaminants in the environment.⁴ Many methods have been investigated to remove Cu(II) from wastewaters such as chemical precipitation, adsorption, ion exchange, membrane separation, reverse osmosis, electrolysis froth flotation and solvent extraction.^5,6 Among these methods, adsorption has increasingly received more attention as it is simple and cost effective. Up to now, various adsorbents have been used to remove Cu(II), such as soil, clay mineral⁷ and resins.^3,8,9 However, all things considered, these materials suffer from either high cost or low adsorption capacity, which limit their value for further study.

Carbonaceous materials are promising adsorbents as they are facile and environmentally friendly.¹⁰ But standard methods for the manufacture of activated carbons usually require very harsh conditions and have several drawbacks.¹¹ What’s worse, the adsorption capacity of activated carbon needs to be further enhanced.¹² Graphene oxide¹³ and carbon nanotubes¹⁴ have shown high adsorption capacities due to their large specific surface areas and numerous surface functional groups,^15–17 while their application still remains at the theory stage.¹⁸

Here, a new type of carbonaceous nanocomposite was synthesized easily via a one-step aqueous hydrothermal carbonization route using cheap glucose in the presence of acrylic acid and a small amount of graphene oxide. The weight ratio of glucose to graphene oxide was fixed at 24 to lower the cost and acquire a large specific surface area.¹⁹ And acrylic acid added to the reaction could offer abundant functional groups and adsorption sites. The prepared materials, containing rich acidic groups, were not only inexpensive and easy to obtain, but also exhibited an excellent adsorption affinity and high maximum adsorption capacity for Cu(II). In order to assess the adsorption capacity of the as-prepared materials and understand the adsorption mechanism, batch experiments, including the effect of pH, ionic strength, complex anions and temperature, were performed and combined with X-ray photoelectron spectroscopy.²⁰ Considering the presence of many natural organic compounds in the real environment, both humic acid and fulvic acid were evaluated.²¹ In addition, the presence of 1-butyl-3-methylimidazolium chloride ([Bmim][Cl]) was also studied, as it is the most used water soluble ionic liquid²² and also a potential contaminant.²³

Experimental

Material preparation

Firstly, graphene oxide was prepared by the improved Hummers method.^24,25 To obtain materials with a high degree of functionality, an acrylic acid monomer was added to the reaction mixture of glucose (3.6 g) and graphene oxide (0.15 g) at different amounts: 0, 3.6, 7.2 and 10.8 mL (mole ratios of acrylic acid to glucose were 0, 2.5, 5 and 7.5). Then deionized water was added to reach a total volume of 70 mL. And the final mixture was sealed into the Teflon inlet in an autoclave, and was hydrothermally treated at 198 °C for 16 h. Then, the materials were washed several times with deionized water and vacuum-dried overnight at 50 °C. The final products obtained were ultrasonically dispersed in deionized water as a stock suspension (1 mg mL⁻¹).

Characterization

The prepared materials were collected and characterized by element content analysis (EC), scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS), Raman spectroscopy, Fourier transform infrared (FTIR) spectroscopy, powder X-ray diffraction (PXRD) and potentiometric titration. EC was performed with a Vario EL (ELEMENTAR). SEM was performed using a MIRA 3 XMU (Tecan). TEM was performed using a Tecnai-G2-F30 Field Emission Transmission Electron Microscope (FEI Corporation). Raman spectra from 500 to 4000 cm⁻¹ were collected on an inVia-Reflex Raman scope using a 632.8 nm He–Ne laser (Renishaw). FTIR spectra were recorded from 400 to 4000 cm⁻¹ on a NEXUS 6705-DX 170SX spectrometer (Nicolet Instrument Corporation). XPS spectra were obtained using an ESCALAB 250Xi (Thermo Fisher Scientific). The PXRD patterns were scanned from 5° to 60° on an X’Pert PRO X-ray diffractometer (Panalytical). The potentiometric acid–base titrations were conducted under argon using a DL50 Automatic Titrator (Mettler Toledo) in NaCl as a background electrolyte.

Adsorption experiments

Batch adsorption experiments were carried out at a 200 rpm equivalent shaking rate in 10 mL polyethylene test tubes for 24 h. The stock suspension of adsorbents and the stock solutions of Cu(II) were added in polyethylene test tubes to achieve the desired concentrations of the different components (solid-to-liquid ratio was fixed at 0.2 g L⁻¹). In most experiments, the solution pH was adjusted to 2–10 as measured with a pH-meter (pHS-3C, Shanghai). For cases where the pH was adjusted, a 0.1 mol L⁻¹ NaOH or HCl solution was used to change the initial pH value. A 5 mol L⁻¹ NaCl (NaNO₃ or NaClO₄) background electrolyte solution was used to adjust ionic strength. After each adsorption equilibrium, the mixtures were centrifuged and the concentration of Cu(II) in the supernatant was determined using an Atomic Absorption Spectrometer (AAS, Perkin-Elmer). The adsorption amount was calculated according to the difference between the Cu(II) concentrations before and after adsorption. Adsorption isotherms were obtained by analysing different concentrations of Cu(II) at fixed temperature (298 K) and pH (5.0). At each condition, adsorption experiments were performed in triplicate and then the average was calculated.

Adsorption kinetics fitting model

The pseudo-second-order model^26,27 general form was:
where t (h) is the contact time, Q_t (mg g⁻¹) is the adsorption amount at time t, Q_e (mol g⁻¹) is the adsorption amount at the equilibrium time, and k (g mg⁻¹ h⁻¹)is the pseudo-second-order rate constant.

Adsorption isotherm fitting model

The General Langmuir Freundlich model^28,29 was employed to fit the adsorption data. Its general form was:
where Q_e (mg g⁻¹) is the amount of Cu(II) adsorbed on the adsorbent; Q_m (mg g⁻¹) is the maximum amount of Cu(II) adsorbed per unit mass of the adsorbent; C_e (mg L⁻¹) is the adsorbate concentration at equilibrium; k (L mg⁻¹) is the adsorption affinity coefficient; and α is an indicator of isotherm nonlinearity related to the heterogeneity of sorption sites.

Results and discussion

Characterization results

It has been reported that three-dimensional graphene-based porous materials were synthesized at the bulk scale via hydrothermal carbonization of glucose and graphene oxide.¹⁹ The reported materials showed large specific surface areas, but lacked surface functional groups for adsorption of metal ions. So we added acrylic acid as a comonomer to the reaction mixture of glucose and graphene oxide to provide more surface functional groups. For instance, GA0, GA1, GA2 and GA3 respectively represent 0, 3.6, 7.2 and 10.8 mL of acrylic acid used in the reaction. Fig. 1 shows the morphological images of the materials. Compared to the smooth surface of GA0 (Fig. 1a), large amounts of carbon nanodots (∼2.8 nm, ESI Fig. S1†) were found to be deposited on the surfaces of GA1, GA2 and GA3 (Fig. 1b–d). And the SEM images (Fig. 1e and f) of the materials confirm that the surface of the materials became rough after adding acrylic acid. The morphological characterization results proved that adding acrylic acid successfully enhanced the surface functional groups.

Fig. 1 TEM images of GA0 (a), GA1 (b), GA2 (c) and GA3 (d); SEM images of GA0 (e) and GA1 (f).

The element content analysis (Table 1) revealed that the oxygen content increased with increasing acrylic acid volume in the reaction. Compared with that of the precursors, an oxygen content of 38.6% seemed to be very high for the final product. The FTIR spectra of the products were recorded (Fig. 2a), and the following functional groups were identified in all samples: O–H stretching vibrations (3431 cm⁻¹), C=O stretching vibrations (1705 cm⁻¹), C=C from sp² carbon (1627 cm⁻¹), C–O vibrations (1406 cm⁻¹), and O–H bending vibrations (1203 cm⁻¹).²⁵ XPS (Fig. S4†) and Raman spectroscopy (Fig. 2b) were used to further analyze the carbon hybridization and oxygen-containing functional group state. The characteristic peaks of the D-band (1360 cm⁻¹) and G-band (1601 cm⁻¹)²⁴ were clear in the Raman spectra. And the results that the I_D/I_G of GA0, GA1, GA2 and GA3 gradually reduced indicated that the sp² carbon content increased with increasing acrylic acid volume in the reaction, because the carbon atoms in the acrylic acid molecule demonstrated sp² hybridization. The powder X-ray diffraction (Fig. 2c) confirmed the broad (002) peak of carbon.

Table 1 Selected properties of the materials

Material	Element content				ID/IG	O1s/C1s	Hs (mmol g^−1)
	C (%)	O (%)	H (%)	O/C
GA0	68.15	27.67	4.18	0.41	1.49	0.31	1.48
GA1	63.48	32.34	4.18	0.51	0.56	0.54	5.37
GA2	59.77	35.19	5.03	0.59	0.20	0.74	6.54
GA3	55.97	38.62	5.41	0.69	0.17	0.76	7.57

---

Fig. 2 FTIR (a), Raman spectra (b), XRD (c) and surface charge densities (ΔQ^H) (d) of GA0, GA1, GA2 and GA3.

Potentiometric titrations were performed under argon atmosphere to characterize the surface charge densities (ΔQ^H, mol g⁻¹) and the total concentration of surface acidic groups per solid weight (H_s, mmol g⁻¹) of material.³⁰ ΔQ^H was determined by subtracting the titration curve of the background electrolyte solution (blank) from that of the material suspension. In general, a suspension of material in 50 mL 0.1 mol L⁻¹ NaCl solution was titrated with a standard NaOH solution (0.05 mol L⁻¹) up to pH ∼ 10 at room temperature. Fig. 2d shows that the pH_PZC (pH of point of zero charge) depended strongly on oxygen content.

(C_a − C_b)_susp = [H⁺] − [OH⁻] + ΔQ_solid + ΔQ_blank

(C_a − C_b)_blank = [H⁺] − [OH⁻] + ΔQ_blank

ΔQ^H = V/m[(C_a – C_b)_susp − (C_a – C_b)_blank]

where C_a and C_b (mol L⁻¹) are the concentrations of acid and base added, respectively; ΔQ_blank (mol L⁻¹) represents consumption or release of H⁺ by side reactions; ΔQ_solid (mol L⁻¹) and ΔQ^H (mol g⁻¹) represents the proton excess of the material in different units, respectively; V (L) is the volume of aqueous solution and m (g) is the mass of material.

H_s calculated from the two equivalence points on the Gran plot (V_eb1 and V_eb2, ESI S2 and Table S1†) is defined by the following equation:

H_s = (V_eb2 − V_eb1)C_b/m_s

The H_s values calculated from the titration curves are listed in Table 1. It was clear that the higher the oxygen content, the higher the H_s value, while the relationship was non-linear. These characterization results indicated that carbonaceous nanocomposites with different degrees of functionality were successfully synthesized (detailed data are summarized in Table 1). Then GA1 was chosen to systematically study the adsorption mechanism.

Kinetics

Fig. 3a shows the adsorption kinetics of Cu(II) on GA1. The adsorption amount increased quickly in the first 1 h, and then slowed down until the adsorption process reached equilibrium, which indicated that the adsorption mechanism was mainly chemical adsorption or surface complexation rather than physical sorption.³ After simulating several adsorption kinetics models, the pseudo-second-order model was found to be the most suitable (fitting parameter in Table S2†) with the correlation coefficient (R²) of 0.998. Although the adsorption process was fast, the shaking time in the following batch experiments was selected as 24 h to achieve complete equilibrium.

Fig. 3 Adsorption of Cu(II) on GA1 as a function of contact time. (a) Plot of Q_t vs. t for Cu(II) adsorption on GA1. (b) Plot of t/Q_t vs. t for the pseudo-second-order model. T = 298 K, pH = 5, initial C_Cu(II) = 0.2 mmol L⁻¹ and I_NaCl = 0.1 mol L⁻¹.

Effect of pH

The pH of a solution influences the surface chemical properties of a suspended material and the species of metal ions.³¹ So it is important to study the effect of pH on adsorption behaviors. Fig. 4a shows the adsorption of Cu(II) on GA1 as a function of pH. The sharp increase of the adsorption percentage over pH 4 was attributed to the deprotonation of the surface acidic groups which had a strong affinity for the positively charged Cu(II) species. According to the Cu(II) speciation distribution as a function of pH value (Fig. 4b), Cu(II) began to form a precipitate at pH 6 if no Cu(II) was adsorbed, while most Cu(II) was adsorbed by GA1 before pH 6, which proved that the adsorption of Cu(II) was not due to the surface precipitation of Cu(OH)_2(s).²¹ In order to completely eliminate the influence of Cu(II) precipitation, the adsorption isotherms of Cu(II) on the materials were obtained at pH 5.0.

Fig. 4 Adsorption of Cu(II) on GA1 as a function of pH (a), T = 298 K, C_Cu(II) initial = 0.2 mmol L⁻¹ and I_NaCl = 0.1 mol L⁻¹. Concentration of Cu(II) species as a function of pH value (b), T = 298 K C_Cu(II) = 0.2 mmol L⁻¹ and I_NaCl = 0.1 mol L⁻¹.

Effect of ionic strength and anion

The ionic strength and anion are important factors to study the adsorption mechanism, so the effect of NaCl, NaNO₃ and NaClO₄ concentration on the adsorption percentage of Cu(II) on GA1 is shown in Fig. 5a. The adsorption was inhibited gradually by NaCl due to strong coordination of Cl⁻ reducing the actual concentration of Cu(II); while the adsorption was promoted by NaNO₃ and NaClO₄, suggesting that the oxideacid anion took part in surface complexation. Comparing the adsorption isotherms of Cu(II) on GA1 in the presence of 0.1 mol L⁻¹ NaCl, NaNO₃ or NaClO₄ (Fig. 5b, fitting parameters in Table S3†), similar results were observed. The results suggested that the materials were favorable for use with wastewater containing the oxideacid anion.

Fig. 5 The effect of NaCl, NaNO₃ and NaClO₄ concentration on the adsorption percentage of Cu(II) on GA1 at pH = 5 and T = 298 K (a). Adsorption isotherms of Cu(II) on GA1 in 0.1 mol L⁻¹ NaCl, NaNO₃ and NaClO₄, pH = 5 and T = 298 K (b).

Effect of temperature

Temperature is also a key factor that affects ion adsorption. Fig. 6 shows the adsorption isotherms of Cu(II) on GA1 at 298, 313 and 328 K (fitting parameters in Table S4†). Temperature showed a slightly positive influence on the adsorption behaviors, which indicated that the adsorption of Cu(II) on GA1 is an endothermic process. Some researchers reported that Cu(II) ions had to be denuded of their hydration sheaths to some extent, and this dehydration process needs energy.^3,32 The materials showed stable adsorption behaviors against temperature.

Fig. 6 Adsorption isotherms of Cu(II) on GA1 at pH = 5, I_NaCl = 0.1 mol L⁻¹, T = 298, 313 and 328 K.

Effect of the presence of organic compounds

In the natural environment, there are many coexistent organic compounds. So the effect of the presence of humic acid, fulvic acid and [Bmim][Cl] were studied, and the results are shown in Fig. 7 (fitting parameters in Table S5†). The results showed that the presence of humic acid and fulvic acid greatly promoted the adsorption capacity, while, on the contrary, the presence of [Bmim][Cl] slightly inhibited the adsorption behaviors. Those organic compounds were easily adsorbed by the materials through π–π interaction. And the negatively charged humic acid and fulvic acid²¹ were coordinated with the Cu(II), forming a Cu–O bond, and also further negatively charged the materials, which significantly improved the adsorption capacity of the positively charged Cu(II). However, although the adsorbed [Bmim]⁺ could form a strong Cu–N bond which promoted the adsorption behaviors, the positively charged [Bmim]⁺ changed the surface charge properties of the materials and inhibited further adsorption of positively charged Cu(II). The results clearly indicated that numerous natural organic compounds would promote the decontamination process.

Fig. 7 Adsorption isotherms of Cu(II) on GA1 in the presence of 20 mg L⁻¹ humic acid, fulvic acid and [Bmim][Cl] at pH = 5, T = 298 K and I_NaCl = 0.1 mol L⁻¹.

Effect of degree of functionality

Four kinds of materials with different degrees of functionality were prepared. The adsorption isotherms of Cu(II) on GA0, GA1, GA2 and GA3 at pH 5, 0.1 mol L⁻¹ NaCl and T = 298 K are shown in Fig. 8. The maximum adsorption amounts (Q_m) of GA0, GA1, GA2 and GA3 were 10.2, 97.2, 122.6 and 146.1 mg g⁻¹, respectively (fitting parameters in Table 2). And it seemed to be clear that the higher the degree of functionality, the higher the adsorption capacity. Modeling the relationship between Q_m and oxygen content or H_s (Fig. 9 and S5†) showed that the Q_m was linearly dependent on H_s (R² = 0.99995). The results indicated that the adsorption behavior was related to the surface acidic groups, not the oxygen content, because some inner oxygen atoms at the hydrothermal carbon layer couldn’t adsorb Cu(II). And previous reports also claimed that the adsorption of metal ions on a series of carbons was found to be related to the type and concentration of the surface functional groups.³³

Fig. 8 Adsorption isotherms of Cu(II) on GA0, GA1, GA2 and GA3 at pH = 5, T = 298 K and I_NaCl = 0.1 mol L⁻¹.

Table 2 The fitting parameters for the general Langmuir Freundlich model of Cu(II) adsorption isotherms

Material	Qm (mg g^−1)	k	α	R^2
GA0	10.2	0.096	1.85	0.73
GA1	97.2	0.332	2.47	0.98
GA2	122.6	0.358	2.24	0.97
GA3	146.1	0.314	2.46	0.96

---

Fig. 9 Relationship between Q_m and H_s.

To further evaluate our result, XPS studies of GA1 before and after Cu(II) adsorption were conducted. The XPS results (Fig. 10) show that the relative content of C–O–C/C–OH was reduced, that of C=O/COO increased, and the binding energy (B.E.) shifted after adsorption which suggested that oxygen atoms were the main adsorption sites and Cu²⁺–O coordinate bonds were formed in the adsorption process.⁵ And the Cu2p XPS also clearly manifested that Cu(II) was adsorbed on the surface of the material.

Fig. 10 XPS of GA1 before adsorption (GA1) and after adsorption (GA1-Cu).

Compared to the adsorption capacities of other adsorbents (Table 3), such as modified activated carbon,¹² graphene oxide¹⁵ and pecan nutshell,³⁴ the adsorption capacity of the carbonaceous nanocomposites in this work was as high as 146.1 mg g⁻¹, which was the highest.

Table 3 Comparison of the Cu(II) adsorption capacities of different adsorbents

Adsorbents	Experimental conditions	Qm (mg g^−1)	Reference
Graphene oxide aerogel	pH = 6.3, T = 298 K	19.7	16
Oxidized multi-walled CNTs	pH = 4, T = 298 K	29.7	14
Modified activated carbon	pH = 5, T = 298 K	38	11
H2SO4/KMnO4-modified CNTs	pH = 6, T = 300 K	38.6	8
CMCD-MNPs	pH = 6, T = 298 K	47.2	5
Graphene oxide	pH = 5, T = 308 K	74.9	15
Pecan nutshell	pH = 5.5, T = 298 K	91.2	34
Carbonaceous nanocomposites	pH = 5, T = 298 K	146.1	This work

---

Competitive adsorption

The competitive adsorption of multi-metal ions on the materials was studied to explore their potential for practical applications. The initial solution was prepared with 0.8 mmol L⁻¹ of each of Cu(II), Sr(II), Zn(II), Mg(II), Ca(II), Cd(II), K(I) and Na(I). After adsorption to the carbonaceous nanocomposites, the concentrations in the supernatant were determined using ICP-AES, and the adsorption percentages are shown in Fig. 11. The adsorptions of Cu(II), Sr(II), and Zn(II) on the materials were higher than those of the other five ions, and the adsorption affinity (∼48%) of the materials for Cu(II) was the strongest when compared with Sr(II) (∼26%) and Zn(II) (∼14%), which meant that the carbonaceous nanocomposites had a high potentiality for selectively removing Cu(II).

Fig. 11 The competitive adsorption of multi-metal ions, C[metal ions]_initial = 0.8 mmol L⁻¹, pH = 5, T = 298 K.

Conclusion

Carbonaceous nanocomposites with a high degree of functionality were synthesized via one-pot hydrothermal carbonization of cheap glucose in the presence of acrylic acid and a small amount of graphene oxide. The materials exhibited an excellent adsorption affinity for Cu(II) and the maximum adsorption capacity was as high as 146.1 mg g⁻¹, which was much higher than any previous reports and was linearly dependent on surface acidic groups. The adsorption kinetics showed that the adsorption behavior was fast and well fitted with the pseudo-second-order kinetic model. And, more remarkably, the presence of the oxideacid anion and natural organic compounds could promote the adsorption behaviors. A schematic diagram elucidating the whole process is illustrated in Fig. 12. Overall, the cheap carbonaceous nanocomposites in this work exhibited a high potential for efficient decontamination of copper ions.

Fig. 12 A schematic diagram elucidating the whole process. The materials were synthesized via one-pot hydrothermal carbonization. Then the materials were used to efficiently remove Cu(II) from aqueous solution.

Acknowledgements

This project is supported by National Science Foundation of China J1210001, 21327801 and 21405165. The authors thank Professor AFM Rahman and Doctor Zhang Rui for modification of the manuscript.

References

1. J. Duruibe, M. Ogwuegbu and J. Egwurugwu, Int. J. Phys. Sci., 2007, 2, 112–118.
2. L. Järup, Br. Med. Bull., 2003, 68, 167–182.
3. X. Wu, D. Zhao and S. Yang, Desalination, 2011, 269, 84–91.
4. C. Yuwei and W. Jianlong, Chem. Eng. J., 2011, 168, 286–292.
5. A. Z. M. Badruddoza, A. S. H. Tay, P. Y. Tan, K. Hidajat and M. S. Uddin, J. Hazard. Mater., 2011, 185, 1177–1186.
6. F. Fu and Q. Wang, J. Environ. Manage., 2011, 92, 407–418.
7. R. Zhang, C. Chen, J. Li and X. Wang, Appl. Surf. Sci., 2015, 349, 129–137.
8. C.-Y. Kuo, Desalination, 2009, 249, 781–785.
9. J. He, Y. Lu and G. Luo, Chem. Eng. J., 2014, 244, 202–208.
10. X. Guo, C. Wang, M. Chen, J. Wang and J. Zheng, J. Power Sources, 2012, 214, 107–112.
11. R. Demir-Cakan, N. Baccile, M. Antonietti and M.-M. Titirici, Chem. Mater., 2009, 21, 484–490.
12. L. Monser and N. Adhoum, Sep. Purif. Technol., 2002, 26, 137–146.
13. R. K. Upadhyay, N. Soin and S. S. Roy, RSC Adv., 2014, 4, 3823–3851.
14. J. Wang, Z. Li, S. Li, W. Qi, P. Liu, F. Liu, Y. Ye, L. Wu, L. Wang and W. Wu, PLoS One, 2013, 8, 1016–1020.
15. X. Ren, J. Li, X. Tan and X. Wang, Dalton Trans., 2013, 42, 5266–5274.
16. G. Huang, W. Wang, X. Mi, W. Xie, Y. Liu and J. Gao, Carbon, 2012, 50, 4856–4864.
17. M. Yusuf, F. M. Elfghi, S. A. Zaidi, E. C. Abdullah and M. A. Khan, RSC Adv., 2015, 5, 50392–50420.
18. G. Zhao, T. Wen, X. Yang, S. Yang, J. Liao, J. Hu, D. Shao and X. Wang, Dalton Trans., 2012, 41, 6182–6188.
19. L. Zhang, F. Zhang, X. Yang, G. Long, Y. Wu, T. Zhang, K. Leng, Y. Huang, Y. Ma and A. Yu, Sci. Rep., 2013, 3.
20. G. Sheng, Y. Li, X. Yang, X. Ren, S. Yang, J. Hu and X. Wang, RSC Adv., 2012, 2, 12400–12407.
21. X. Tan, M. Fang, J. Li, Y. Lu and X. Wang, J. Hazard. Mater., 2009, 168, 458–465.
22. D. J. Gorman-Lewis and J. B. Fein, Environ. Sci. Technol., 2004, 38, 2491–2495.
23. B. Jastorff, K. Mölter, P. Behrend, U. Bottin-Weber, J. Filser, A. Heimers, B. Ondruschka, J. Ranke, M. Schaefer and H. Schröder, Green Chem., 2005, 7, 362–372.
24. D. C. Marcano, D. V. Kosynkin, J. M. Berlin, A. Sinitskii, Z. Sun, A. Slesarev, L. B. Alemany, W. Lu and J. M. Tour, ACS Nano, 2010, 4, 4806–4814.
25. T. Longlong, Z. Xin, Q. Wei, L. Dan, J. Qiang, L. Jin, Y. Yuanlv, L. Zhan and W. Wangsuo, RSC Adv., 2014, 4, 58536–58545.
26. L. Wang, C. Cheng, S. Tapas, J. Lei, M. Matsuoka, J. Zhang and F. Zhang, J. Mater. Chem. A, 2015, 3, 13357–13364.
27. S. W. Zhang, J. X. Li, T. Wen, J. Z. Xu and X. K. Wang, RSC Adv., 2013, 3, 2754–2764.
28. M. N. Carvalho, C. A. M. de Abreu, M. Benachour, D. C. S. Sales, O. S. Baraúna and M. A. D. Motta Sobrinho, Adsorpt. Sci. Technol., 2012, 30, 691–699.
29. G. P. Jeppu and T. P. Clement, J. Contam. Hydrol., 2012, 129–130, 46–53.
30. Q. H. Fan, P. Li, W. S. Wu, D. Q. Pan, K. F. Ding, Y. Lu, T. Yu and J. Xu, Sci. China: Chem., 2011, 54, 1138–1147.
31. A. Schierz and H. Zänker, Environ. Pollut., 2009, 157, 1088–1094.
32. M. Wang, H. Jiang and Z. C. Wang, J. Therm. Anal. Calorim., 2006, 85, 751–754.
33. B. Xiao and K. M. Thomas, Langmuir, 2005, 21, 3892–3902.
34. J. C. P. Vaghetti, E. C. Lima, B. Royer, B. M. da Cunha, N. F. Cardoso, J. L. Brasil and S. L. P. Dias, J. Hazard. Mater., 2009, 162, 270–280.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra16451f

‡ These authors contributed equally to this work.

---