Zero valent iron/poly(amidoxime) adsorbent for the separation and reduction of U(VI)

Abstract

A magnetic nanocomposite of zero valent iron (ZVI)/poly(amidoxime) (PAO) was prepared by polymerization of acrylonitrile on the ZVI surface, followed by hydroxylamine (NH₂OH) treatment to convert the cyano group (–CN) into an amidoxime group (AO). The characterization results showed that PAO-stabilized ZVI magnetic composite (denoted as ZVI/PAO) was prepared successfully. ZVI/PAO was used as an adsorbent for the separation of U(VI) from solutions. The effect of environmental conditions on U(VI) adsorption onto ZVI/PAO was studied in detail. U(VI) adsorption on the ZVI/PAO surface reached equilibrium within 3 h, and the adsorption process is well described by the pseudo-second-order kinetics model. Adsorption isotherms of U(VI) on the ZVI/PAO surface can be well fitted by the Langmuir model, and the maximum adsorption capacity was calculated to be 206 mg g⁻¹ at T = 298 K and pH = 5.0. Experimental results highlighted the application of ZVI/PAO as an adsorbent for the efficient separation of U(VI) and reduction of U(VI) to U(IV) from aqueous solutions.

---

Introduction

Abundant radioactive contaminants were formed during the past decades with the fast development and widespread application of nuclear technology and energy, which is a crucial worldwide environmental problem that affects human health. Meanwhile, in order to cover increasing energy demand, a large number of nuclear power plants have and will be built to generate electrical power. Uranium (U(VI)), a representative long-life radionuclide, is a very essential element in nuclear techniques.^1–3 Therein, the effective enrichment and separation of U(VI) from solution has become a very pressing issue to public health and sound development of nuclear techniques.

Magnetic adsorbents have attracted great interest for the separation of contaminants from solution because they can be magnetically recovered from solution after use.^4,5 Among them, ZVI^6–9 is widely recognized as an important remediation material for radioactive contaminants due to its high adsorption capability and magnetic property. However, the easy aggregation and sedimentation of ZVI nanoparticles in solution greatly limit its real application. Therefore, significant advances have been made on ZVI stabilization.^10,11 Surface modification of ZVI is a practical method to improve its stabilization and dispersion properties in solution. PAO-based adsorbents have attracted wide interest for the enrichment and separation of U(VI) ions from solution due to their attractive properties, such as excellent adsorption capacity, high selectivity, rapid adsorption rate, easy handling, and recyclability.^2,12,13 Based on the reasons mentioned above, ZVI/PAO would present an excellent sorption ability for U(VI) ions, and the U-laden ZVI/PAO can be magnetically recovered on a large scale. However, to the best of our knowledge, few researchers have ever seriously evaluated the physicochemical properties of ZVI/PAO, and its application as an adsorbent in U(VI) contamination management.

In this study, ZVI/PAO was prepared by polymerization of acrylonitrile on the ZVI surface, followed by hydroxylamine (NH₂OH) treatment. The as-synthesized ZVI/PAO was applied as an adsorbent to enrich and separate U(VI) ions from solution. The results revealed that ZVI/PAO had excellent performance in the separation of U(VI) from solution. This work highlights the potential application of ZVI/PAO in U(VI) contamination management.

Results and discussion

Characterization of ZVI/PAO composites

To observe the morphology of ZVI/PAO, SEM images of ZVI and ZVI/PAO were measured (Fig. 1). SEM images demonstrate that ZVI (Fig. 1A and B) and ZVI/PAO (Fig. 1C and D) were spherical-like in shape (150–250 nm), and presented wrinkled and crumpled surfaces, which can provide convenient migration and diffusion channels for U(VI) ions into the interior of ZVI/PAO, and are thereby in favour of U(VI) adsorption when ZVI/PAO was applied as an adsorbent for the enrichment and separation of U(VI) ions from aqueous solution. The surfaces of ZVI/PAO are smoothed and covered by organic materials after being modified with PAO, indicating the successful modification of PAO on ZVI surfaces.

Fig. 1 SEM images of ZVI (A and B) and ZVI/PAO (C and D).

The powder XRD pattern of ZVI/PAO is shown in Fig. 2A. Two obvious peaks at 2θ = 44.7° and 65.2° correspond to the (110) and (200) diffractions of α-Fe (JCPDS no. 06-0696), respectively.^14–17 The result indicates that ZVI is prepared successfully. In addition, small impurity peaks of iron oxide (2θ = 35.6° and 53.7°) are also observed in the XRD pattern of ZVI/PAO, indicating that some iron oxide is formed on the ZVI/PAO surface.¹⁴ Many researchers have found the formation of iron oxide in laboratory-made ZVI, and they suggested a core (ZVI)–shell (iron oxide) structure.^15–17

Fig. 2 XRD patterns (A), TGA curves (B), magnetization curves (C), and FT-IR spectra (D) of ZVI, PAO, and ZVI/PAO.

TGA curve characterization is used to calculate the weight percent of PAO in ZVI/PAO. The TGA curve of ZVI/PAO shows a simple thermal degradation process (Fig. 2B). The 6.9% weight loss at ∼27–90 °C can be assigned to the loss of adsorbed moisture. The ∼45.7% weight loss at ∼86–500 °C is attributed to the decomposition of PAO. Considering the complete degradation and decomposition of PAO at ∼650 °C in N₂ conditions,¹⁸ the mass ratio of PAO to ZVI and the weight percent of PAO in dry ZVI/PAO are 0.964 : 1 and 49.1%, respectively. This result coincides well with the experimental conditions and provides further evidence that PAO is successfully modified on the ZVI surface.

The magnetic property of ZVI/PAO is also studied (Fig. 2C), which is critical for the magnetic separation of ZVI/PAO from solution in potential applications. The hysteresis loops of ZVI/PAO at room temperature reveal the soft magnetic features of ZVI/PAO. These materials typically exhibit high permeability and low coercivity.¹⁹ The saturation magnetization (M_s) of ZVI/PAO is measured to be 63.1 emu g⁻¹ at a magnetic field of 0 ± 30 kOe and at room temperature. The inset in Fig. 2C shows that ZVI/PAO particles move quickly toward the magnet, and a clear solution is obtained within minutes. It suggests the potential application of ZVI/PAO as a magnetic adsorbent in the enrichment and separation of U(VI) ions from aqueous solution, followed by magnetic separation after application in real applications.

The functional groups on the ZVI/PAO surface can be classified by FT-IR spectroscopy analysis and are shown in Fig. 2D. ZVI does not show any typical FT-IR peaks. ZVI/PAO shows characteristic FT-IR peaks related to PAO, which include (1) –OH stretching vibrations at 3417, 1656, and 799 cm⁻¹,^20–22 (2) N–H anti-symmetric vibrations at 3224 and 1130 cm⁻¹,²³ and (3) CH₂ stretching at 2930, 2856, and 1352 cm⁻¹.^21,24 The sound peak at 897 cm⁻¹ is related to stretching of the N–O bond in PAO.^25,26 These results clearly confirm that ZVI/PAO is successfully prepared. Meanwhile, the characteristic FT-IR peak at ∼2240 cm⁻¹ related to the vibration of –CN^20,23–25 is not detected, which proves the high conversion of PAN to PAO.^22,24

The functional groups on the ZVI/PAO surface are also studied by the XPS technique. The successful preparation of ZVI/PAO is confirmed by the XPS survey spectrum in Fig. 3A. The peaks at ∼711, ∼400, and ∼285 eV can be assigned to Fe 2p, N 1s, and C 1s, which confirms the successful modification of PAO on the ZVI surface. The XPS Fe 2p spectrum (Fig. 3B) is deconvoluted into Fe(0), Fe(II) and Fe(III), which are centered at 707.87, 710.05, and 711.46 eV,^27–29 respectively (Table 1). It confirms that a shell layer of iron oxide is formed on the ZVI surface. The XPS N 1s spectrum (Fig. 3C) is deconvoluted into N≡C–CH, H₂N–C=NOH, and positively charged N atoms (N⁺), which are centered at 399.44, 400.16, and 401.56 eV, respectively (Table 1). The amidoximation degree (DA) of PAO is obtained from the conversion of PAN into PAO.³⁰ According to the area ratio of the peak at 400.16 eV to the peak at 399.44 eV, the DA value is calculated to be ∼0.98. The high DA value of ZVI/PAO confirms the high conversion of PAN to PAO. The XPS C 1s spectrum of ZVI/PAO (Fig. 3D) can be deconvoluted into CH₂–CH and CH–C(NH₂)=NOH, which are centered at 284.67 and 285.84 eV, respectively. The area ratio of CH₂–CH to CH–C(NH₂)=NOH is calculated to be ∼1 : 2.1, which is in good agreement with the theoretical calculation.³⁰

Fig. 3 XPS survey (A), Fe 2p (B), N 1s (C), and C 1s (D) spectra of ZVI/PAO.

Table 1 Curve fitting results of the XPS Fe 2p, N 1s, and C 1s spectra of ZVI/PAO

	Peak	BE^a (eV)	FWHM^b (eV)	%
a Binding energy.b Full width at half-maximum.
Fe 2p	Fe(0)	707.87	1.24	2.60
	Fe(II)	710.05	1.88	39.1
	Fe(III)	711.46	3.39	58.3
N 1 s	N≡C	399.44	1.47	47.9
	H2N–C=NOH	400.16	1.76	47.0
	N^+	401.56	0.96	5.09
C 1 s	CH2–CH	284.67	1.23	32.3
	CH–C(NH2)=NOH	285.84	1.93	67.7

---

Adsorption study

The adsorption kinetics can well reflect the adsorption efficiency of ZVI/PAO for U(VI) ions. To investigate the adsorption dynamics of ZVI/PAO in potential applications, the effect of contact time on the adsorption of U(VI) on the ZVI/PAO surface is studied and shown in Fig. 4A. U(VI) adsorption on ZVI/PAO and PAO increases quickly at the initial 3 h, and then keeps a high level. Although the amount of U(VI) adsorbed on ZVI is much lower than that on PAO and on ZVI/PAO under the same experimental conditions, almost all the U(VI) is adsorbed by ZVI after one month contact time (data not shown). Reductive immobilization of soluble U(VI) ions to insoluble U(IV) species by ZVI is a combination of processes: the corrosion of ZVI, the electron transfer from ZVI to U(VI), and the adsorption of U(VI) ions on the formed iron oxide,^6,31,32 which usually needs a very long time to accomplish the reduction process because the shell of iron oxide can effectively prevent the core of ZVI from reacting with U(VI) ions. Therefore, the adsorption of U(VI) on the ZVI/PAO surface is a combined process of the fast adsorption process of U(VI) ions by PAO, and slow reduction process of soluble U(VI) ions to insoluble U(IV) species by ZVI. The adsorption of U(VI) on ZVI/PAO was accompanied by the release of Fe(II)/Fe(III) into aqueous solution. The concentration of Fe(II)/Fe(III) increased with the increasing contact time, and fewer Fe(II)/Fe(III) ions were detected in the ZVI/PAO suspensions (Fig. 4A). It confirms that the shell of iron oxide and PAO can prevent the core of ZVI from reacting with U(VI).

Fig. 4 Effect of contact time (A), adsorption isotherms (B), effect of pH and ionic strength (C), the relative distribution of 2.0 × 10⁻⁴ mol L⁻¹ U(VI) (D), and effect of temperature (E and F) on U(VI) adsorption on ZVI/PAO, PAO, and ZVI. (A) T = 298 ± 1 K, C_{[U(VI)]initial} = 2.0 × 10⁻⁴ mol L⁻¹, m/V = 0.20 g L⁻¹, I = 0.1 mol L⁻¹ NaCl, pH = 5.0 ± 0.1. (B) T = 298 ± 1 K, contact time = 24 h, m/V = 0.20 g L⁻¹, I = 0.1 mol L⁻¹ NaCl, pH = 5.0 ± 0.1. (C) T = 298 ± 1 K, contact time = 24 h, C_{[U(VI)]initial} = 2.0 × 10⁻⁴ mol L⁻¹, m/V= 0.20 g L⁻¹. (E and F) contact time = 24 h, m/V = 0.20 g L⁻¹, I = 0.1 mol L⁻¹ NaCl, pH = 5.0 ± 0.1.

A rapid adsorption rate favours the application of the ZVI/PAO composite in the efficient elimination of U(VI) ions from aqueous solution. The typical pseudo-first-order kinetics model (q_t = q_e × (1 − exp(−k₁t)), where q_e (mg g⁻¹) and q_t (mg g⁻¹) represent the experimental and maximum adsorption capacity, respectively, and k₁ (h⁻¹) is the rate constant) and pseudo-second-order kinetics model (t/q_t = 1/k₂q_e² + t/q_e, where k₂ (g mg⁻¹ h) is the rate constant) are used to fit the experimental data (Table 2). The adsorption process of U(VI) ions on ZVI is better described by the pseudo-first-order kinetics model, whereas the adsorption process of U(VI) on ZVI/PAO and PAO is better described by the pseudo-second-order kinetics model. The latter is usually used to describe a chemisorption process.³³ Therefore, the adsorption of U(VI) on the ZVI/PAO surface is mainly controlled by a chemisorption mechanism rather than by physical adsorption.

Table 2 Comparison of the pseudo-first-order and pseudo-second-order parameters for U(VI) adsorption on ZVI/PAO

	Pseudo-first-order			Pseudo-second-order
	qe (mg g^−1)	K1 (h^−1)	R^2	qe (mg g^−1)	K2 (g mg^−1 h)	R^2
ZVI	28.9	0.035	0.999	106	0.0011	0.987
PAO	129	2.88	0.973	129	0.118	0.999
ZVI/PAO	156	2.55	0.900	158	0.0448	0.999

---

Adsorption isotherms are usually used to estimate the adsorption capacity of an adsorbent. As depicted in Fig. 4B, the adsorption data of U(VI) on ZVI/PAO are obviously higher than those of U(VI) on ZVI and on PAO under the same experimental conditions. To evaluate the adsorption capability of ZVI/PAO for U(VI) ions, the adsorption data are analyzed by a classical Langmuir model (C_s = bC_smaxC_e/(1 + bC_e), where C_e and C_s are the U(VI) concentrations in solution and on the adsorbent surface, respectively, and C_smax and b are the maximum adsorption capacity and the related constant, respectively) and the Freundlich model (C_s = K_FC_eⁿ, where K_F and n present the adsorption capacity and adsorption intensity at a specific temperature, respectively). The Langmuir model assumes that adsorption sites on the adsorbent surface are homogeneous, and the sorbate forms only a monolayer. Unlike the Langmuir model, the Freundlich model is applicable to adsorption behaviour occurring on heterogeneous surfaces. It is believed that the adsorbed amount is closely related to the solute concentration. According to the R² data in Table 3, the ZVI/PAO surfaces are homogeneous for U(VI) adsorption. The C_smax value of U(VI) on the ZVI/PAO surface is calculated to be 206 mg g⁻¹ at pH 5.0. Furthermore, ZVI/PAO kept its physicochemical properties and adsorption capability for U(VI) after being stored for one year. It suggests the excellent stability of ZVI/PAO, which ensures the safe application of ZVI/PAO in real work.

Table 3 Parameters calculated from the Langmuir and Freundlich models for U(VI) adsorption on ZVI, PAO, and the ZVI/PAO composite at pH = 5.0 and T = 298 ± 1 K

	Langmuir model			Freundlich model
	Csmax (mg g^−1)	b (L mg^−1)	R^2	K (mg g^−1)	1/n	R^2
ZVI	45.8	0.167	0.996	12.4	0.334	0.948
PAO	188	0.0926	0.993	32.0	0.473	0.946
ZVI/PAO	206	0.205	0.996	65.0	0.284	0.932

---

Fig. 4C shows the effect of pH on U(VI) adsorption onto the ZVI/PAO surface. The adsorption of U(VI) increases significantly with the pH increasing in the pH range of 1.0–7.0, and then keeps a high level at pH > 7.0. At pH ∼7.0, ZVI/PAO can effectively adsorb >95% U(VI) ions from aqueous solution. The pH-dependent adsorption behaviour suggests that U(VI) adsorption is controlled by the distributions of U(VI) species and the protonation–deprotonation reactions of ZVI/PAO. The relative distributions of U(VI) species (2.0 × 10⁻⁴ mol L⁻¹) in aqueous solution are calculated from the hydrolysis constants of U(VI)³ and are given in Fig. 4D, which are fairly dependent on pH. U(VI) mainly exists as the prevailing species of free uranyl (UO₂²⁺) up to pH 4.5, and then U(VI) hydrolysis complexes and U(VI) multinuclear hydroxide complexes become to be the important existing forms of U(VI). Meanwhile, PAO is protonated at low pH values,^34,35 which can hinder U(VI) adsorption on the ZVI/PAO surface, which can explain the poor U(VI) adsorption at low pH values. As the pH value increases, the protonation degree of adsorption sites and the electrostatic repulsion are gradually decreased, which can enhance U(VI) adsorption on ZVI/PAO at high pH values. The influence of ionic strength is performed at different NaCl concentrations (e.g. 0.01, 0.10 and 0.50 mol L⁻¹). As shown in Fig. 4C, U(VI) adsorption decreases with increasing ionic strength. The ionic strength dependence of U(VI) adsorption onto the ZVI/PAO surface might be explained by the fact that ionic strength can affect the electrical double layer of the ZVI/PAO surface and the activity of U(VI) species, which limits the migration of U(VI) from solution to ZVI/PAO surfaces, and thereby affects U(VI) adsorption.

Thermodynamic parameters can provide important information about the energetic change involved in the adsorption process. Three different temperatures (e.g. 278, 298, and 318 K) are selected to study the impaction of environmental temperature on U(VI) adsorption on ZVI/PAO and to calculate the related thermodynamic parameters (Fig. 4E). U(VI) adsorption on ZVI/PAO increases at higher temperatures. The related distribution coefficient (K_d, Fig. 4F) and thermodynamic parameters (Table 4) are obtained from the temperature-dependent adsorption data.^36,37 The positive value of the standard enthalpy change (ΔH⁰) indicates an endothermic adsorption process of U(VI) on the ZVI/PAO surface, which is supported by the increased U(VI) adsorption with increasing temperature. The Gibbs free energy change (ΔG⁰) can reveal the affinity and the related driving forces of adsorption. The negative values of ΔG⁰ and the positive value of entropy change (ΔS⁰) prove that the adsorption process of U(VI) on the ZVI/PAO surface is spontaneous and thermodynamically favourable.

Table 4 Thermodynamic parameters and the parameters calculated from the Langmuir models for U(VI) adsorption on ZVI/PAO

	Langmuir model			Thermodynamic parameters
	Csmax (mg g^−1)	b (L mg^−1)	R^2	ΔG^0 (kJ mol^−1)	ΔH^0 (kJ mol^−1)	ΔS^0 (J mol^−1 K)
278 K	181	0.235	0.995	−24.2	3.03	99.8
298 K	206	0.205	0.996	−26.2
318 K	225	0.192	0.995	−28.2

---

In order to explore the enrichment mechanism of U(VI) on ZVI/PAO, the U-laden ZVI/PAO is characterized by XRD, XPS, and SEM images, and elemental mapping characterizations. The formation of iron oxide on U-laden ZVI/PAO is confirmed by XRD patterns. As shown in Fig. 5A, after contact of 24 h, the relative intensity of iron oxide is significantly increased whereas that of ZVI is decreased. It confirms that ZVI is partly transferred into iron oxide. The reduction of soluble U(VI) to insoluble U(IV) is coupled with the oxidation of ZVI,^6,38,39 which is further evidenced and confirmed by the results of XPS characterization. The XPS U 4f spectrum of ZVI/PAO after U(VI) adsorption (Fig. 5B) is quantitatively deconvoluted into U(VI) (381.9 and 392.7 eV) and U(IV) (380.7 and 391.6 eV).⁴⁰ The results reveal that the adsorbed U(VI) is partly reduced to U(IV) by ZVI. The adsorbed U on the ZVI/PAO surface is further evidenced by elemental mapping characterizations (Fig. 5C). The SEM image indicates that ZVI/PAO retains its stacking morphology after U(VI) adsorption, which reveals the excellent stability of ZVI/PAO. Elemental mapping images indicate that C, O, and U are homogeneously dispersed on the ZVI surface, which further confirms the successful synthesis of ZVI/PAO and its excellent adsorption capacity for U(VI).

Fig. 5 XRD patterns (A), XPS U 4f spectrum (B), and SEM images and elemental mapping (C) of U-laden ZVI/PAO.

Conclusions

The results suggest the successful preparation of the ZVI/PAO composite. U(VI) adsorption on the ZVI/PAO surface is accomplished within an hour of contact time, and can be fitted by the pseudo-second-order kinetics model. The ZVI/PAO composite possesses excellent adsorption capacity to U(VI) in solution with the advantage of a simple magnetic separation process, which highlights the application of ZVI/PAO as a candidate adsorbent for separation of U(VI) from aqueous solution. It then reduces soluble U(VI) species to insoluble U(IV) materials to decrease the environmental mobility of U(VI).

Experimental section

Preparation of ZVI/PAO

The preparation process of the ZVI/PAO nanoparticles consists of the following three steps: (1) preparation of ZVI nanoparticles. 13.9 g FeSO₄·7H₂O was dispersed in 100 mL degassed Milli-Q water, and then reduced to metallic ZVI nanoparticles by the incremental addition of 100 mL NaBH₄ solution (30.3 g L⁻¹). After stirring for 1 h, the suspension was separated by a common magnet and rinsed with degassed Milli-Q water thoroughly. Thus ZVI was obtained. (2) Polymerization of acrylonitrile on the ZVI surface. The obtained ZVI nanoparticles and 2.80 g acrylonitrile were dispersed in 100 mL degassed Milli-Q water in an ice-water bath. Under vigorous mechanical stirring and Ar conditions, 100 mL NaOH/urea/H₂O solution (7 : 12 : 81 by weight) was added drop-wise into the suspension to induce the polymerization of acrylonitrile.⁴¹ The reaction mixture was kept in the ice-water bath for 8 h. The resultant ZVI/poly(acrylonitrile) (denoted as ZVI/PAN) materials were rinsed with degassed Milli-Q water, and then dried at 60 °C for 24 h under Ar conditions. (3) NH₂OH treatment. The resultant ZVI/PAN nanoparticles were reacted with NH₂OH (100 mL, 66.5 g L⁻¹) in a methanol–water solution (80 : 20, v/v) at 70 °C for 3 h under vigorous mechanical stirring and Ar conditions.² After NH₂OH treatment, the samples were rinsed with the methanol–water solution, and dried at 60 °C for 24 h under Ar conditions. Thus the ZVI/PAO nanoparticles were obtained.

Characterization

The physicochemical properties of ZVI/PAO were studied by Fourier transform infrared spectroscopy (FT-IR), X-ray photoelectron spectroscopy (XPS), powder X-ray diffraction (XRD), thermogravimetric analysis (TGA), a vibrating sample magnetometer (VSM), scanning electron microscopy (SEM), and elemental mapping in detail. The FT-IR measurement was mounted on a Perkin-Elmer 100 spectrometer. XPS spectroscopy was carried out by an ESCALab220i-XL surface microanalysis system (VG Scientific) equipped with an Al Kα (hν = 1486.6 eV) source. XRD characterization was performed using X-ray diffraction (Rigaku D/max) with Cu Kα radiation (λ = 0.1541 nm). Surface charging effects were corrected with the C 1s peak at 284.4 eV as a reference. TGA measurement was performed on a thermogravimetric analyzer (Shimadzu TGA-50) at a N₂ flow rate of 50 mL min⁻¹ with a heating rate of 10 °C min⁻¹. The VSM curve was obtained on a model 155 VSM in magnetic fields up to ± 30 kOe at room temperature. SEM imaging and elemental mapping were performed on JSM-6320F FE-SEM (JEOL).

Batch adsorption experiments

The batch adsorption experiment was carried out in 10 mL centrifuge tubes in an anoxic glovebox. After placing the saturated CaO–Ca(OH)₂ absorber in, the chamber was purged with high purity N₂ to minimize the effect of CO₂ and O₂. The suspension of adsorbent and NaCl was shaken for 24 h at 298 ± 1 K to guarantee the pre-equilibrium of the adsorbent and background ions. After U(VI) was added, the pH values were adjusted by highly concentrated HCl or NaOH, and then shaken for the desired time. The suspensions were centrifuged at 18 000 rpm at 298 ± 1 K for 20 min (BECKMAN COULTER 64R, USA). The finial concentration of U(VI) in the supernatant was detected by the arsenazo III spectrophotometric method (650 nm). All experimental data were the average of triplicate determinations, and the data relative errors were <5%.

Acknowledgements

Financial support from NSAF (1U1530131), National Natural Science Foundation of China (91326110, 21225730, 91326202 and 21577032), the Radiochemistry 909 Project in China Academy of Engineering Physics, the Science and Technology Development Foundation of China Academy of Engineering Physics (2014B0301034), and Anhui Provincial Natural Science Foundation (1508085MB29) are acknowledged.

Notes and references

1. D. Shao, G. Hou, J. Li, T. Wen, X. Ren and X. Wang, Chem. Eng. J., 2014, 255, 604–612.
2. D. Shao, J. Li and X. Wang, Sci. China: Chem., 2014, 57, 1449–1458.
3. D. Shao, Z. Jiang, X. Wang, J. Li and Y. Meng, J. Phys. Chem. B, 2009, 113, 860–864.
4. D. Shao, C. Chen and X. Wang, Chem. Eng. J., 2012, 185–186, 144–150.
5. D. Shao, J. Hu, C. Chen, G. Sheng, X. Ren and X. Wang, J. Phys. Chem. C, 2010, 114, 21524–21530.
6. R. A. Crane, M. Dickinson, I. C. Popescu and T. B. Scott, Water Res., 2011, 45, 2931–2942.
7. G. Sheng, Y. Tang, W. Linghu, L. Wang, J. Li, H. Li, X. Wang and Y. Huang, Appl. Catal., B, 2016, 192, 268–276.
8. G. Sheng, P. Yang, Y. Tang, Q. Hu, H. Li, X. Ren, B. Hu, X. Wang and Y. Huang, Appl. Catal., B, 2016, 193, 189–197.
9. G. Sheng, X. Shao, Y. Li, J. Li, H. Dong, W. Cheng, X. Gao and Y. Huang, J. Phys. Chem. A, 2014, 118, 2952–2958.
10. R. L. Johnson, R. J. O’Brien, J. T. Nurmi and P. G. Tratnyek, Environ. Sci. Technol., 2009, 43, 5455–5460.
11. H. H. Tseng, J. G. Su and C. Liang, J. Hazard. Mater., 2011, 192, 500–506.
12. Y. Zhao, J. Li, S. Zhang and D. Shao, RSC Adv., 2013, 3, 18952–18959.
13. Y. Zhao, J. Li, L. Zhao, S. Zhang, Y. Huang, X. Wu and X. Wang, Chem. Eng. J., 2014, 235, 275–283.
14. M. Z. Kassaee, E. Motamedi, A. Mikhak and R. Rahnemaie, Chem. Eng. J., 2011, 166, 490–495.
15. Z. Pang, M. Yan, X. Jia, Z. Wang and J. Chen, J. Environ. Sci., 2014, 26, 483–491.
16. Z. Chen, X. Jin, Z. Chen, M. Megharaj and R. Naidu, J. Colloid Interface Sci., 2011, 363, 601–607.
17. R. L. Frost, Y. Xi and H. He, J. Colloid Interface Sci., 2010, 341, 153–161.
18. S. M. Badawy, H. H. Sokker and A. M. Dessouki, J. Appl. Polym. Sci., 2006, 99, 1180–1187.
19. E. J. Kim, J. H. Kim, A. M. Azad and Y. S. Chang, ACS Appl. Mater. Interfaces, 2011, 3, 1457–1462.
20. P. Lv, Y. Bin, Y. Li, R. Chen, X. Wang and B. Zhao, Polymer, 2009, 50, 5675–5680.
21. G. Tian, J. Geng, Y. Jin, C. Wang, S. Li, Z. Chen, H. Wang, Y. Zhao and S. Li, J. Hazard. Mater., 2011, 190, 442–450.
22. S. El-Khouly, Y. Takahashi, A. A. Saafan, E. Kenawy and Y. A. Hafiz, J. Appl. Polym. Sci., 2011, 120, 866–873.
23. H. H. Abdel-Razik and E. R. Kenawy, J. Appl. Polym. Sci., 2012, 125, 1136–1145.
24. X. Liu, H. Liu, H. Ma, C. Cao, M. Yu, Z. Wang, B. Deng, M. Wang and J. Li, Ind. Eng. Chem. Res., 2012, 51, 15089–15095.
25. J. Chen, R. Qu, Y. Zhang, C. Sun, C. Wang, C. Ji, P. Yin, H. Chen and Y. Niu, Chem. Eng. J., 2012, 209, 235–244.
26. S. Pal, V. Ramachandhran, S. Prabhakar, P. K. Tewari and M. Sudersanan, J. Macromol. Sci., 2006, 43, 735–747.
27. J. Shi, S. Yi, H. He, C. Long and A. Li, Chem. Eng. J., 2013, 230, 166–171.
28. R. Yang, Y. Wang, M. Li and Y. Hong, ACS Sustainable Chem. Eng., 2014, 2, 1270–1279.
29. Y. Sun, X. Li, J. Cao, W. Zhang and H. P. Wang, Adv. Colloid Interface Sci., 2006, 120, 47–56.
30. Z. J. Yu, E. T. Kang and K. G. Neoh, Langmuir, 2002, 18, 10221–10230.
31. O. Riba, T. B. Scott, K. V. Ragnarsdottir and G. C. Allen, Geochim. Cosmochim. Acta, 2008, 72, 4047–4057.
32. B. Gu, L. Liang, M. J. Dickey, X. Yin and S. Dai, Environ. Sci. Technol., 1998, 32, 3366–3373.
33. S. Das, A. K. Pandey, A. A. Athawale and V. K. Manchanda, J. Phys. Chem. B, 2009, 113, 6328–6335.
34. M. Atta, A. A. H. Abdel-Rahman, M. F. Hamza, I. E. El Aassy and F. Y. Ahmed, J. Dispersion Sci. Technol., 2012, 33, 490–496.
35. D. Bilba, N. Bilba and G. Moroi, Sep. Sci. Technol., 2007, 42, 171–184.
36. D. Shao, X. Ren, J. Hu, Y. Chen and X. Wang, Colloids Surf., A, 2010, 360, 74–84.
37. P. Li, Z. Yin, J. Lin, Q. Jin, Y. Du, Q. Fan and W. Wu, Environ. Sci.: Processes Impacts, 2014, 16, 2278–2287.
38. S. Luo, P. Qin, J. Shao, L. Peng, Q. Zeng and J. Gu, Chem. Eng. J., 2013, 223, 1–7.
39. S. R. Kanel, B. Manning, L. Charlet and H. Choi, Environ. Sci. Technol., 2005, 39, 1291–1298.
40. E. S. Ilton and P. S. Bagus, Surf. Interface Anal., 2011, 43, 1549–1560.
41. W. Li, R. Liu, H. Kang, Y. Sun, F. Dong and Y. Huang, Polym. Chem., 2013, 4, 2556–2563.

---