Efficient removal of Pb(II) ions using manganese oxides: the role of crystal structure

Abstract

Manganese oxides have been proven to be promising adsorbents to capture Pb(II) from wastewaters. In nature, MnO₂ can be found in different crystalline structures, while the effect of crystal structure on their adsorption performance remains unclear. In this study, five manganese oxides with different crystallographic phases, α-, β-, γ-, δ-, and λ-MnO₂ were prepared and characterized by X-ray diffraction (XRD), transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS), N₂ adsorption–desorption, Fourier transform infrared (FT-IR) spectroscopy and zeta potential measurements. The adsorptive removal of aqueous Pb(II) was investigated using these manganese oxides as adsorbents. The results showed that the adsorption capacities of manganese oxides for Pb(II) varied with BET surface area and crystalline structure, following the order of δ-MnO₂ > α-MnO₂ > λ-MnO₂ > γ-MnO₂ > β-MnO₂. δ-MnO₂ displayed the highest capacity for Pb(II), and the adsorption was scarcely influenced by the presence of the coexisting Na⁺ cation. The surface complexation model was used to describe the Pb(II) adsorption on the MnO₂ adsorbents. In a column adsorption test δ-MnO₂ was capable of continuously treating 25 000 bed volumes synthetic wastewater stream with an influent concentration of 20 mg L⁻¹ Pb(II) and an effluent concentration below 0.5 mg L⁻¹. This work indicates that δ-MnO₂ has great potential to be used as an effective adsorbent for Pb(II) removal.

---

1. Introduction

Water pollution by Pb(II) is a serious environmental problem. Wastewaters from various industrial activities usually contain considerable amounts of Pb(II) which can be a severe threat to public health.¹ Toxicity studies have demonstrated that Pb(II) can be accumulated in human body because of its high volatility and non-biodegradability, causing neurological, reproductive and immunological pathologies.^2,3 Therefore, a simple and cost-effective treatment technique to remove Pb(II) from the environment is recognized as a top priority by the World Health Organization (WHO).⁴

Adsorption is considered as an effective and economic technology to remove Pb(II) from water prior to its discharge into the natural environment. A variety of materials have been explored for adsorptive removal of Pb(II) from water, including carbon materials,⁵ clay minerals^6,7 and metal oxides (e.g. Fe₂O₃, Al₂O₃ and MnO₂),⁸ and MnO₂ was found to be the one of the most active species for Pb(II) removal.⁹ Wang et al.¹⁰ prepared layered birnessite type MnO₂ and found that the adsorbent was effective for the removal of Pb(II) at pH 4.5. In parallel, Maliyekkal et al. synthesized cellulose–manganese oxide hybrid material for adsorptive removal of Pb(II) in water.¹¹ It should be emphasized that manganese oxides are suitable for Pb(II) adsorption because of the high specific surface area, low isoelectric point (IEP) and negatively charged surface over a wide pH range.¹² Tonkin et al.¹³ reported that Pb(II) could be effectively adsorbed on δ-MnO₂ within the pH range of 2–9. Moreover, manganese oxides have attracted special attention because of their outstanding structural flexibility, low-cost and environmental friendliness, which have been widely applied in adsorbents,⁹ catalysts¹⁴ and cathodes.¹⁵

Manganese oxides are widely used in adsorption of removing Pb(II) from aqueous solutions. However, according to some recent studies, the adsorption affinity of MnO₂ towards Pb(II) was highly related to the structures and surface properties of MnO₂. Luo et al.¹⁶ studied adsorptive removal of lithium by 3D-MnO₂ from wastewater and found that 3D-MnO₂ with high degree of porosity and large pores exhibited a higher adsorption capacity than 1D-MnO₂. Kozlova et al.¹⁷ prepared complex manganese oxides with different sizes of structural tunnels, and concluded that the tunnel size was the determining factor for the sorption of heavy metal cations. In addition, Korshin et al.¹⁸ and Abdullah et al.¹⁹ found that heavy metal ions could be complexed at different adsorption sites on birnessite and α-MnO₂ using EXAFS spectroscopy. Wang et al.¹⁰ studied the adsorption mechanisms of Pb(II) onto varied birnessite type MnO₂ and found that the adsorption capacities were largely controlled by the Mn average oxidation states and type of sorption sites on the MnO₂ surface. Notably, MnO₂ exists in five main crystallographic forms in nature: α-, β-, γ-, δ-, and λ- types, as the basic structure unit, [MnO₆] octahedron, links in different conditions.²⁰ The crystalline structures are generally believed to be responsible for their chemical and physical properties.²¹ The influence of crystal forms on MnO₂ catalytic performance has been extensively studied.^22,23 Different crystallographic MnO₂ materials were also prepared for pollutants adsorption, and MnO₂ adsorbents showed varying adsorption performance depending on the surface area and crystalline structure. Zhang et al. investigated the effect of the crystallographic structure of MnO₂ on Li⁺ adsorption, and found that MnO₂ ion sieve with λ-crystallographic structure exhibited higher capacity for Li(I) selective adsorption.²⁴ Kanungo and Parida prepared MnO₂ adsorbents with different crystallographic forms (α-, β-, γ- and δ-MnO₂), and adsorption of Cu(II) on those adsorbents were studied. They found the layered δ-MnO₂ showed the highest adsorption capacity and the adsorption capacity was directly related to BET surface area.²⁵ To gain a better understanding of the effect of MnO₂ structure on adsorption behaviors, the adsorption comparison of Pb(II) on α-, β-, γ-, δ-, and λ-MnO₂ were made in our experiments. However, to the best of our knowledge, no attempt has been made.

In this study, α-, β-, γ-, δ-, and λ-MnO₂ were prepared and characterized by X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), transmission electron microscopy (TEM), Fourier transform infrared (FT-IR), N₂ adsorption/desorption and zeta potential measurements. The adsorption properties of the adsorbents toward Pb(II) in aqueous solution were systematically examined for adsorption capacity and effect of water chemistry. The influence of crystal forms of MnO₂ on their adsorption performance was investigated in conjunction with important physicochemical parameters. Column adsorption tests were also carried out to elucidate the applicability of δ-MnO₂ in Pb(II) removal from synthetic wastewater effluent.

2. Experimental section

2.1 Materials and reagents

Chemicals used in the experiments were of analytical grade and were used without further purification. All chemicals including manganese sulfate (MnSO₄·H₂O), ammonium persulfate ((NH₄)₂S₂O₈), potassium hypermanganate (KMnO₄), lead nitrate (Pb(NO₃)₂), sodium nitrate (NaNO₃), magnesium nitrate (Mg(NO₃)₂), calcium nitrate (Ca(NO₃)₂) were purchased from Nanjing Chemical Reagents Co., Ltd. Stock solutions of Pb(II) was prepared by dissolving appropriate amount of Pb(NO₃)₂ in deionized (DI) water.

2.2 Preparation of manganese oxides

The α-MnO₂, β-MnO₂, and γ-MnO₂ were prepared by the hydrothermal method according to the literature.²⁰ In brief, 2.4 g of MnSO₄·H₂O and appropriate amount of (NH₄)₂S₂O₈ (5.3 g for α-MnO₂, and 3.7 g for β-MnO₂ and γ-MnO₂) were put into DI water at room temperature to form a homogeneous solution, which was then transferred into a 40 mL Teflon-lined stainless steel autoclave and heated for 12 h. The hydrothermal temperature was set at 90 °C for γ-MnO₂, and 140 °C for α-MnO₂ and β-MnO₂. After the reaction was completed, the resulting solid product was filtered and washed with DI water, drying at 110 °C for 12 h.

To prepare δ-MnO₂, 1.8 g of KMnO₄ was dissolved in 100 mL DI water, then 30 mL H₂O₂ (30%) was added dropwise to the solution under fast stirring. The solution pH was kept at 7.0 ± 0.2 using 0.1 mol L⁻¹ HNO₃ and/or KOH. After continuously stirring for another 30 min, the suspension was aged at room temperature for 3 h. Then the aged precipitate was filtered and washed with DI water and dried at 110 °C for 12 h.

The λ-MnO₂ was prepared using LiMn₂O₄ as the precursor. Four grams of KMnO₄, 1.5 g of NaCl and 4.5 g of LiCl were grinded thoroughly and transferred to ceramic crucible, followed by a heat-treatment at 600 °C for 5 h.²⁶ The resulting solid product was added in 200 mL 0.1 mol L⁻¹ HNO₃ solution under fast stirring for 12 h. After the acid treatment, the suspension was filtered and the solid product was washed with DI water and dried at 110 °C for 12 h.

2.3 Characterization of manganese oxides

Powder XRD patterns of the samples were collected on a Rigaku D/max-RA powder diffraction meter with Cu Kα radiation (Rigaku, Japan). TEM images of the samples were collected on a Hitachi H-800 transmission electron microscope (Hitachi, Japan). The specific surface areas of the adsorbents were measured by N₂ adsorption/desorption after pretreated at 200 °C under vacuum (1.33 Pa) for 4 h on a Micrometrics ASAP 2020 analyzer (Micromeritics ASAP 2020, USA) at −196 °C (77 K) and were calculated by the Brunauer–Emmett–Teller (BET) method. FT-IR spectra of the samples were recorded on a Nicolet Nexus 870 FT-IR spectrometer (Nicolet, USA) with the KBr pellet technique. An XPS spectrometer (Perkin Elmer PHI 550 ESCA/SAM, USA) equipped with a monochromatized Al Kα X-ray source (hv = 1486.6 eV) and a hemispherical electron analyzer was used to analyze the surface properties of the samples. The C 1s peak (284.6 eV) was used for the calibration of binding energy values. Surface zeta potentials (ζ) of the materials were measured using a zeta potential analyzer (Zeta PALS, Brookhaven Instruments Co., USA). Briefly, 0.01 g of materials were dispersed in 100 mL of 1 mmol L⁻¹ NaCl with varied pH adjusted using 0.1 mol L⁻¹ HCl or NaOH, and the suspension was allowed to settle for 24 h before zeta-potential measurement.

2.4 Adsorption kinetics

To examine the adsorption kinetic of Pb(II) to the adsorbent, 0.25 g of adsorbent was added into a 500 mL round-bottom flasks containing 500 mL of Pb(II) solution with varied initial concentrations under vigorous magnetic stirring. The initial Pb(II) concentrations were 45.0 mg L⁻¹ for α-, β-, γ- and λ-MnO₂, and 150.0 mg L⁻¹ for δ-MnO₂. The samples were taken from the flasks at preset time intervals and the adsorbent was removed by fast filtration. The concentration of Pb(II) in the filtrate was determined by AAS, and the adsorption amount was calculated according to the following equation:where q_t (mg g⁻¹) is the adsorption amount at time t, C₀ (mg L⁻¹) is the initial concentration of Pb(II) in solution, C_t (mg L⁻¹) is Pb(II) concentration at time t, V (L) is the solution volume, and M (g) is the mass of the adsorbent.

2.5 Adsorption isotherms

Adsorption isotherms of Pb(II) on the adsorbents were obtained from batch experiments by placing 20.0 mg of adsorbent in 40 mL polytetrafluoroethylene-lined screw cap glass tubules receiving 40 mL of Pb(NO₃)₂ aqueous solutions with varied initial concentrations. The pH of the aqueous solutions was adjusted to 4.0 ± 0.2 using 0.1 mol L⁻¹ HNO₃ or NaOH. Then, the samples were shaken at room temperature for 24 h in an orbital shaker. The time period was sufficient to reach apparent adsorption equilibrium based on preliminarily determined adsorption kinetics. After shaking, the sample was filtrated using a 0.45 μm filter, and Pb(II) concentration in the was filtrate determined by AAS. The adsorbed amount of Pb(II) at equilibrium was calculated according to the following equation:where q_e (mg g⁻¹) and C_e (mg L⁻¹) are the adsorbed concentration and aqueous concentration of Pb(II) at equilibrium, respectively, C₀ (mg L⁻¹) is the initial Pb(II) concentration, V (L) is the solution volume, and M (g) is the adsorbent mass.

The impact of ionic strength and pH on Pb(II) adsorption was tested in the presence of NaNO₃ solutions with varied concentrations (0.001, 0.01, and 0.10 mol L⁻¹) in a pH range of 2.5–10.0. The initial Pb(II) concentrations were 45.0 mg L⁻¹ for α-, β-, γ- and λ-MnO₂, and 150.0 mg L⁻¹ for δ-MnO₂. The initial pH of Pb(II) solution was adjusted using 0.1 mol L⁻¹ HNO₃ or NaOH, and the pH of all samples was measured after adsorption equilibrium. All adsorption experiments were conducted in duplicate.

2.6 Fixed-bed column adsorption

The experiments of column adsorption were carried out in a polyethylene column (1.78 mm internal diameter, 130 mm in length) equipped at room temperature. The wet δ-MnO₂ (bed volume = 2.0 mL, M = 3.27 g) or silica sand was packed within the columns before operation. A control of only silica sand was also tested and the results revealed that the silica sand did not adsorb Pb(II) (data not shown). A synthetic influent solution containing Pb(II) (20.0 mg L⁻¹), Na(I) (100.0 mg L⁻¹), Mg(II) (100.0 mg L⁻¹) and Ca(II) (150.0 mg L⁻¹) was pumped upflow through the column at flow rate of 15.0 mL h⁻¹ using a Lange-580 pump (Baoding, China). The effluent was collected regularly by a BSZ-100 programmable fraction collector (Shanghai, China) and measured by AAS. As for the regeneration test, a mixed solution containing 0.1 mol L⁻¹ HNO₃ and 5 wt% Ca(NO₃)₂ was used as the regenerant and the flow rate was 7.5 mL h⁻¹.

3. Results and discussion

3.1 Characterization of adsorbents

The XRD patterns of varied MnO₂ adsorbents are compiled in Fig. 1. For α-MnO₂, diffraction peaks were identified with 2θ at 22.5°, 28.7°, 37.5°, 42.0°, 49.9°, 56.6°, 59.9°, 65.3° and 68.9°, indicative of the presence of manganese oxide with a tetragonal structure.²⁰ As for β-MnO₂, diffraction peaks were observed with 2θ at 28.6°, 37.3°, 40.9°, 42.8°, 46.1°, 56.7°, 59.3° and 64.9°, assigned to rutile structure.²⁰ The XRD pattern of γ-MnO₂ displayed four distinct diffraction peaks at 22.2°, 37.6°, 42.5° and 56.2°, corresponding to manganese oxide resulted from a random intergrowth of blocks of the rutile-type structural units within the Ramsdellite structure.²⁰ Two low intensity and broad diffraction peaks at 37.0° and 67.5° were observed for δ-MnO₂, characteristic of hexagonal randomly-stacked and single phased birnessite-type MnO₂.²⁷ The diffraction peaks of λ-MnO₂ were identified with 2θ at 36.5°, 38.2°, 44.4°, 48.8°, 58.9°, 64.6°, 65.6° and 67.9°, indicative of cubic spinel structure for λ-MnO₂.²⁶ The XRD results clearly revealed the successful preparation of varied MnO₂ crystallographic forms.

Fig. 1 XRD patterns of synthetic MnO₂ samples.

The TEM images of the five MnO₂ adsorbents are compared in Fig. 2 α-MnO₂ (Fig. 2a) was shown to be ribbon-like nanowires with diameters around 10 nm and lengths ranging between 0.06 and 0.4 μm. β-MnO₂ (Fig. 2b) and γ-MnO₂ (Fig. 2c) were observed to be nanorods with average diameters of 100 and 20 nm, respectively. As for δ-MnO₂ (Fig. 2d), the layered structure was clearly identified, which was commonly formed by MnO₂ nanosheets organizing in layer-by-layer structure for multilayer systems. The λ-MnO₂ (Fig. 2e) had a cubic structure with a size about 20 nm, which was consistent with the XRD results. The BET surface areas were 97.5, 14.2, 71.4, 270.8 and 36.6 for α-, β-, γ-, δ-, and λ-MnO₂, respectively. The δ-MnO₂ with a layered structure exhibited a highest BET surface area, which was likely because the layer spacing of δ-MnO₂ was large enough to host N₂ molecules. As for α-, β-, γ- and λ-MnO₂, N₂ molecules were mainly adsorbed on the external surface.²³

Fig. 2 TEM images of (a) α-MnO₂, (b) β-MnO₂, (c) γ-MnO₂, (d) δ-MnO₂, and (e) λ-MnO₂.

The XPS spectra of the adsorbents in the Mn 2p_3/2 and O 1s regions are compared in Fig. 3a and b, respectively. As can be seen in Fig. 3a, the Mn 2p_3/2 spectrums were asymmetrical and could be divided into three peaks at 639.9, 641.2 and 642.2 eV, assigned to Mn²⁺, Mn³⁺ and Mn⁴⁺, respectively.^10,28 To obtain the relative quantity of Mn²⁺, Mn³⁺ and Mn⁴⁺ in the adsorbents, the Mn 2p_3/2 spectrums were fitted according to the literature,^10,29 and the fitting results are listed in Table 1. The content of Mn³⁺ in α- and β-MnO₂ were higher than others. According to literature, the higher Mn³⁺ content led to fewer cationic vacancies because the Mn³⁺ site binding might involve additional complexes when compared to Mn⁴⁺.¹⁰ The O 1s spectrums (Fig. 3b) of the adsorbents were highly asymmetrical and were divided into two peaks at 528.6–529.0 eV and 530.9–531.0 eV, assigned to oxygen containing species of lattice oxygen and adsorbed oxygen, respectively.³⁰ The contents of lattice oxygen and adsorbed oxygen in the adsorbents are listed in Table 1. As for δ-MnO₂, the content of adsorbed oxygen was the highest among the adsorbents, revealing an abundance of surface hydroxyl groups.

Fig. 3 (a) Mn 2p_3/2 and (b) O 1s XPS spectra of different MnO₂ adsorbents.

Table 1 The results of XPS Mn 2p_3/2 and O 1s multiplets peak fitting

Sorbents	Mn^4+			Mn^3+			Mn^2+			Olatt			OHads
	BE^a (eV)	FWHM^b (eV)	At.^c (%)	BE (eV)	FWHM (eV)	At. (%)	BE (eV)	FWHM (eV)	At. (%)	BE (eV)	FWHM (eV)	At. (%)	BE (eV)	FWHM (eV)	At. (%)
a Binding energy.b The FWHM of all peaks were constrained.c At. represents the percentage of the contribution for each peak to the total number of counts under the Mn 2p3/2 or O 1s peak, and all peaks modeled as 70% Gaussian–30% Lorentzian.
α-MnO2	642.22	2.72	56.1	641.08	1.57	43.9	639.02	1.18	0.0	528.62	1.10	61.7	530.67	3.29	38.3
β-MnO2	642.21	2.93	58.3	641.13	1.62	41.1	639.91	0.91	0.6	528.87	1.15	58.9	530.31	3.40	41.1
γ-MnO2	642.18	2.64	60.1	641.21	1.61	35.2	639.96	0.96	4.7	528.82	1.15	64.0	530.37	3.17	36.0
δ-MnO2	642.29	2.38	58.9	641.35	1.46	38.1	639.99	0.85	3.0	528.97	1.26	55.7	530.10	3.20	44.3
λ-MnO2	642.21	2.78	60.5	641.22	1.63	35.6	639.87	0.90	3.9	528.86	1.43	71.0	530.86	3.53	29.0

---

The functional groups on MnO₂ surface were studied by FT-IR, and the FT-IR spectra of the adsorbents are compiled in Fig. 4. The broad band appearing at around 3400 cm⁻¹ was assigned to O–H stretching and the band at 1610 cm⁻¹ was attributed to O–H bending vibration.³¹ For all MnO₂ adsorbents, the characteristic bands at 400–800 cm⁻¹ were observed, corresponding to the Mn–O lattice vibration.²⁷ For the layered δ-MnO₂ and cubic spinel λ-MnO₂, the bands around 416, 441 and 512 cm⁻¹ were recorded, which is in good agreement with reported FT-IR characteristic bands in literature.²⁷ However, as for the tunnel α-, β- and γ-MnO₂ the position of those characteristic bands displayed some obvious changes. A weak band around 711 cm⁻¹ was observed, corresponding to the manganese oxides with tunnel structure.^32,33 Additionally, the band at 1106 cm⁻¹ was also recorded in the FT-IR spectra of α- and γ-MnO₂, characteristic of the vibration of the Mn³⁺–O bond.²⁷ This observation suggests that some Mn³⁺ occupies the Mn⁴⁺ sites in the frameworks of α- and γ-MnO₂. These results implied that the crystal structure of MnO₂ played a key role in affecting the Mn–O lattice vibration as well as the nature of surface hydroxyl groups.

Fig. 4 FT-IR spectra of MnO₂ adsorbents with different crystallographic forms.

Surface zeta potentials of the varied crystallographic MnO₂ as a function of solution pH are compared in Fig. 5. Within the test pH range, surface zeta potentials of all adsorbents monotonically decreased due to the continuous deprotonation effect of surface hydroxyl groups. The isoelectric points (IEP) of α-, β-, γ-, δ-, and λ-MnO₂ were found to be 2.4, 2.4, 2.0, 1.3 and 2.6, respectively, similar to those reported in literatures.^12,21,26,34 Additionally, negative surface zeta potentials were observed on the five adsorbents over a wide pH range, suggesting the potential for adsorptive removal of Pb(II) in water. Notably, much lower ζ potential was observed in δ-MnO₂ than in others in the test pH range of 1.0–6.0.

Fig. 5 Zeta potential of different MnO₂ adsorbents as a function of pH.

3.2 Adsorption kinetics

Fig. 6a compares the kinetics of Pb(II) adsorption onto the five MnO₂ adsorbents. It was shown that the adsorption of Pb(II) on MnO₂ adsorbents were very fast. The Pb(II) adsorption amounts increased sharply within 20 min and reached equilibrium at about 150 min. The kinetic data were fitted using the pseudo-second-order kinetic model as follow:²¹where q_e (mg g⁻¹) and q_t (mg g⁻¹) are the adsorbed amount at equilibrium and time t (min), respectively, and k (g mg⁻¹ min⁻¹) is the pseudo-second-order rate constant. Fig. 6b displays the linear fitting plots of Pb(II) adsorption on the adsorbents to the pseudo-second-order kinetics. The fitting parameters are summarized in Table 2. High correlation coefficients (R² > 0.99) and the consistence of calculated q_cal values with experimental data indicated that Pb(II) uptake onto the adsorbents could be described by the pseudo-second-order kinetic model and were controlled by chemisorption rather than mass transport.²¹ The adsorption rates of Pb(II) adsorption on the adsorbents were in the order of γ-MnO₂ > α-MnO₂ > β-MnO₂ > λ-MnO₂ > δ-MnO₂, reflecting a lower adsorption kinetic on layered MnO₂ adsorbent.

Fig. 6 (a) Time resolved Pb(II) adsorption on MnO₂ adsorbents, and fitting of Pb(II) adsorption kinetics using (b) pseudo second-order model and (c) Weber–Morris model. The initial Pb(II) concentrations were 45.0 mg L⁻¹ for α-, β-, γ- and λ-MnO₂, and 150.0 mg L⁻¹ for δ-MnO₂.

Table 2 Fitting parameters of Pb(II) adsorption kinetics to MnO₂ adsorbents using pseudo second-order and Weber–Morris model

Sorbents	qexp, mg g^−1	Pseudo second-order			Weber–Morris model
		k, mg g^−1 min^−1	qcal, mg g^−1	R^2	k1, mg g^−1 min^−1/2	C1	R^2	k2, mg g^−1 min^−1/2	C2	R^2
α-MnO2	67.6	4.40 × 10^−3	67.1	0.99	3.63	40.75	0.94
β-MnO2	40.3	3.42 × 10^−3	39.7	0.99	3.78	18.82	0.85
γ-MnO2	47.9	5.94 × 10^−3	47.6	0.99	4.44	18.42	0.90
δ-MnO2	299.2	6.07 × 10^−4	303.0	0.99	36.67	1.65	0.92	20.51	97.27	0.96
λ-MnO2	62.4	3.31 × 10^−3	62.5	0.99	2.59	32.33	0.99

---

To further verify the controlling mechanism involved in the adsorption process, Weber–Morris model was used to fit the kinetic adsorption data. The model is expressed as follows:²¹

q_t = k_it^1/2 + C_i (4)

where k_i (mg g⁻¹ min^−1/2) is the intraparticle diffusion rate constant, and C_i is the intercept.

Fig. 6c presents the plots of q_t versus t^1/2 for adsorption of Pb(II) onto the five adsorbents, and the fitting parameters are listed in Table 2. The q_t–t^1/2 relations clearly exhibited multilinearity, reflecting that the adsorption process appeared to be controlled by more than one step. According to Weber–Morris model, the adsorption process is only controlled by intraparticle diffusion when C_i = 0. For δ-MnO₂, the q_t–t^1/2 plots consisted of three linear portions, and the first linear portion did not pass through the origin, suggesting the involvement of external diffusion.²¹ Therefore, the first linear portion was attributed to the fast lead adsorption onto δ-MnO₂ external surface and the second one was indicative of Pb(II) diffusion into the interlayer space of the adsorbent, and the third one was characteristic of the final equilibrium step. Notably, the majority of Pb(II) adsorption onto δ-MnO₂ was observed in second linear portion and the intraparticle diffusion was slower than the external surface diffusion. In contrast, the Weber–Morris plots of α-, β-, γ- and λ-MnO₂ consisted of two linear portions, in which the first linear portion was Pb(II) diffusion to the sorbents external surfaces and the second one was the final equilibrium stage.

3.3 Adsorption isotherms

The adsorption isotherms of Pb(II) onto various MnO₂ adsorbents are compared in Fig. 7a. For δ-MnO₂, Pb(II) adsorption amount was 242.6 mg g⁻¹ at an equilibrium concentration of 28.2 mg L⁻¹, which was significantly larger than that for α-, β-, γ- and λ-MnO₂ (84.4, 58.9, 62.1 and 61.9 mg g⁻¹, respectively). To analyze the adsorption behavior of the adsorbents, the adsorption isotherms were further modeled using Langmuir and Freundlich models. The Langmuir isotherm model is expressed as follows:where q_m (mg g⁻¹) is the maximum adsorption capacity, and b (L mg⁻¹) is the Langmuir equilibrium constant.

Fig. 7 (a) Pb(II) adsorption isotherms and (b) BET surface area normalized Pb(II) adsorption isotherms on MnO₂ adsorbents.

The Freundlich isotherm model can be described as:

where K_f is Freundlich coefficient that characterizes the adsorption capacity of the adsorbent, and n is the linearity index which reflects the intensity of adsorption.

The fitting parameters are summarized in Table 3. High correlation coefficients (R²) were achieved for the Langmuir and Freundlich models, indicating that Pb(II) adsorption onto the adsorbents could be well described by these two models. The maximum adsorption capacity q_m value and Freundlich coefficient K_f value for δ-MnO₂ were both higher in comparison with the other MnO₂, confirming the largest adsorption capacity of δ-MnO₂. This can be explained by different BET surface area of the five MnO₂ adsorbents. Among the five MnO₂ adsorbents, δ-MnO₂ exhibited the largest BET surface area with abundant hydroxyl groups as observed in XPS and FT-IR spectra.³⁰ Compared with the other three MnO₂ adsorbents, γ- and λ-MnO₂ had larger b values in Langmuir model, suggesting their high affinity for Pb(II). This is likely due to the high contents of Mn⁴⁺ in γ- and λ-MnO₂. The charge imbalance on Mn⁴⁺ sites is much higher than on Mn³⁺ and Mn²⁺ sites and the former sites exhibit higher affinity for metal binding. Thus, structural forms of MnO₂ had great impact on its Pb(II) adsorption performance.

Table 3 Parameters of Langmuir and Freundlich isotherm model for the adsorption of Pb(II) onto sorbents

Sorbents	Langmuir model			Freundlich model
	b (L g^−1)	qm (mg g^−1)	R^2	Kf (mg^1−n L^−1 g^−1)	n	R^2
α-MnO2	1.08	84.0	0.97	45.4	5.08	0.96
β-MnO2	0.55	42.4	0.93	12.5	2.20	0.99
γ-MnO2	12.93	51.6	0.92	34.6	5.35	0.94
δ-MnO2	3.57	200.0	0.92	112.2	3.68	0.87
λ-MnO2	6.92	60.2	0.96	47.5	13.02	0.91

---

The interaction of Pb(II) with surface hydroxyl groups was confirmed by the XPS results of MnO₂ after Pb(II) adsorption. Fig. 8 displays the XPS spectra of Pb 4f and O 1s of the MnO₂ after Pb(II) adsorption. The MnO₂ adsorbents displayed two peaks with Pb 4f_7/2 and Pb 4f_5/2 derived from spin–orbit splitting. The binding energy of Pb 4f_7/2 core level was observed at around 138.6 eV, reflecting the orthorhombic PbO compound and no hydroxides or carbonates precipitation of Pb during adsorption.²¹ As for O 1s spectra, after Pb(II) adsorption the ratio of O_ads decreased in all MnO₂ adsorbents. This was likely due to the complexation interactions between surface –OH and Pb(II). The increase of O_latt ratio in Pb(II)-loaded MnO₂ was attributed to the formation of Pb–O–Mn bond and the higher binding energy of O 1s in Pb(II)-loaded MnO₂ confirmed the covalent of Pb–O–Mn linkage.

Fig. 8 (a) Pb 4f and (b) O 1s XPS spectra of MnO₂ adsorbents after Pb(II) adsorption.

More insight into the structural properties of the adsorbents with varied crystallographic forms could be gained by comparing the surface area normalized adsorption isotherms, and the results are shown in Fig. 7b. Among the tested adsorbents, the normalized adsorption capacity of β-MnO₂ was much higher. This could be well explained by the interface reaction mechanism inferred from adsorbate–adsorbent polyhedral linkage. Previous studies found that the [MnO₆] octahedral structure units of the manganese dioxide were connected with the edge and double-corner linkages, and concluded that Pb(II) was adsorbed at double-corner-sharing (DCS) and double-edge-sharing (DES) sites.³⁵ Notably, β- and λ-MnO₂ is composed of single strands of edge-sharing [MnO₆] octahedral with 1 × 1 tunnels, while α- and γ-MnO₂ are constructed from the chains of 2 × 2, 1 × 1 and 1 × 2 tunnels, respectively.^20,36 Thus, more DCS and DES adsorption sites per unit area were observed on β- and λ-MnO₂ surfaces, making them showed higher surface area normalized adsorption capacities of Pb(II). Additionally, δ-MnO₂ exhibited nearly identical normalized adsorption capacity to that of α- and γ-MnO₂. These results further implied that crystal structure confinement effect was the main factor that affected the adsorption capacity.

The q_m value for δ-MnO₂ (200.0 mg g⁻¹) synthesized in this study was also markedly larger than the values reported for many other adsorbents such as coconut shell carbon (26.5 mg g⁻¹),³⁷ amino-functionalized Fe₃O₄@SiO₂ (76.7 mg g⁻¹),³⁸ montmorillonite (158.5 mg g⁻¹)⁷ and nanoscale organo-functionalized SiO₂–Al₂O₃ (183.0 mg g⁻¹).³⁹ Hence, the δ-MnO₂ exhibited a significantly higher adsorption capacity than these other adsorbents. Additionally, the synthesis cost of δ-MnO₂ was much less than amino-functionalized Fe₃O₄@SiO₂, which was feasible for industrial Pb(II) recycling.

3.4 Effect of pH and ionic strength

The effects of pH and ionic strength on Pb(II) adsorption to the adsorbents were examined, and the results are presented in Fig. 9. For the five adsorbents, Pb(II) adsorption was significantly affected by the solution pH, which increased continuously in the experimental pH range of 2.5–5.5. It should be noted that Pb(II) would precipitate under high pH (e.g. pH > 7.0) conditions. The effect of pH on Pb(II) adsorption on MnO₂ could be well explained by the electrostatic interaction mechanism. Because surface hydroxyl groups were the adsorption sites, the impact of solution pH on Pb(II) adsorption predominantly depending on the protonation–deprotonation transition of surface hydroxyl groups.^13,21 The results of zeta potential showed that the isoelectric points (IEP) of α-, β-, γ-, δ-, and λ-MnO₂ were found to be 2.4, 2.4, 2.0, 1.3 and 2.6, respectively, indicated that MnO₂ was negatively charged at pH values above the IEP and electrostatic attractive interaction was expected at high pH, resulting in increased Pb(II) adsorption. When the solution pH values below IEP, MnO₂ was positively charged and repulsive interaction occurred at low pH between Pb(II) and MnO₂, thus displaying lowered Pb(II) adsorption. The ζ potential analysis showed that MnO₂ surface became more negative charged with increasing solution pH, which would enhance the attraction of Pb(II). The nearly negligible Pb(II) adsorption on the sorbents at pH 2.0 suggested that the saturated adsorbent could be regenerated under acidic conditions.

Fig. 9 Surface binding species for Pb(II) adsorbed on the five MnO₂ adsorbents at different solution pH and ionic strengths.

In NaNO₃ solutions, the formation of aqueous Pb(II) complexes with NO₃⁻ and OH⁻ species might affect the Pb(II) adsorption on MnO₂ as reported in previous work.⁴¹ The aqueous speciation reactions of Pb(II) with the electrolyte anions (OH⁻ and NO₃⁻) were described as follows:^40,41

Pb²⁺ + H₂O = Pb(OH)⁺ + H⁺, log K = −6.19 (7)

Pb²⁺ + 2H₂O = Pb(OH)₂(s) + 2H⁺, log K = −8.15 (8)

Pb²⁺ + NO₃⁻ = PbNO₃⁺, log K = 1.1 (9)

Pb²⁺ + 2NO₃⁻ = Pb(NO₃)₂, log K = 1.4 (10)

The equilibrium protonation constants and complexation constants with OH⁻ and NO₃⁻ were calculated at the ionic strength of 0.01 mol L⁻¹ NaNO₃, and these constants were also used for the higher ionic strengths. The predicted aqueous complexation of Pb(II) in 0.001, 0.01 and 0.1 mol L⁻¹ NaNO₃ solutions was performed using the computer program FITEQL 3.2, and the results are shown in Fig. 9. The amount of Pb(OH)⁺ species increased with the solution pH, while the amount of PbNO₃⁺ species increased with ionic strength. According to the literature, the increase of ionic strength facilitated the formation of the Pb(NO₃)⁺ and Pb(NO₃)₂ species,⁴⁰ whose amounts increased at higher ionic strength at low pH conditions (<5.0). Moreover, the stability of aqueous Pb(II) complexes was enhanced at high ionic strength, which might suppress the adsorption of Pb(II) on the adsorbents. However, due to the high adsorption capacity of α-, γ-, δ- and λ-MnO₂, the PbNO₃⁺ and Pb(OH)⁺ species could also be adsorbed on the negative charged MnO₂ in the tested pH range, suggesting the negligible effect of speciation on Pb(II) adsorption onto α-, γ- and δ-MnO₂. As for β- and λ-MnO₂, the poor Pb(II) adsorption capacity within the pH range of 2.5–5.0 implied that a large amount of PbNO₃⁺ and Pb(OH)⁺ complexes remained in the solutions. The residual PbNO₃⁺ and Pb(OH)⁺ complexes would precipitate at high pH conditions (>6.0), which was in line with the results of pH effect. Additionally, the observed effect of ionic strength on Pb(II) adsorption on MnO₂ was negligible over the pH range of 2–10, indicating that Pb(II) was adsorbed predominantly via inner-sphere complexation. Similar results were also observed by Al Degs et al.⁴² and Han et al.⁴³ This could be ascribed to the fact that the five MnO₂ adsorbents possessed excellent adsorption selectivity for Pb(II) under the interference of Na(I). Moreover, δ-MnO₂ exhibited a high adsorption capacity of Pb(II) with outstanding selectivity under high ionic strength, indicating a promising potential for Pb(II) adsorptive removal from wastewaters.

To verify the efficiency of using δ-MnO₂ to remove Pb(II) from real wastewater, Pb(II) adsorption on δ-MnO₂ in a synthetic wastewater containing Na(I) (100.0 mg L⁻¹), Mg(II) (100.0 mg L⁻¹) and Ca(II) (150.0 mg L⁻¹) was studied, and the results are shown in Fig. 10. It is shown that Pb(II) adsorption on the δ-MnO₂ adsorbent slightly increased with the co-existing ions in synthetic wastewater at pH 4.0, further indicating that Pb(II) adsorption was mainly driven by the strong inner-sphere complexation interactions between Pb(II) and the hydroxyl groups on δ-MnO₂.²¹ The observed effects of coexisting alkaline/earth metal ions on Pb(II) adsorption to δ-MnO₂ in synthetic wastewater clarified the validity of using δ-MnO₂ as an efficient adsorbent for Pb(II) removal from actual wastewater.

Fig. 10 Adsorption isotherms of Pb(II) on δ-MnO₂ in a synthetic wastewater containing 100 mg L⁻¹ of Na(I), 100 mg L⁻¹ Mg(II) and 150 mg L⁻¹ Ca(II).

3.5 Modeling the adsorption of Pb(II) onto MnO₂

A number of previous researchers have used the diffuse double layer surface complexation model (SCM) to predict the adsorption of Pb(II) on manganese oxides at different ionic strengths.^44,45 The SCM assumes that only surface complexation occurs on the charged surfaces, consistent with the adsorption mechanism of Pb(II) adsorption to the five MnO₂ adsorbents. Thus, the SCM was chosen to be used in this study to describe the adsorption of Pb(II) to the surface sites on the five MnO₂ adsorbents. The determination of surface equilibrium constants from experimental data was performed using the computer program FITEQL 3.2. The surface hydroxyl groups, including DCS and DES sites, at the five MnO₂ adsorbents surface were determined according to the literature,⁴⁴ and the results are listed in Table 4. The content of hydroxyl groups on δ-MnO₂ was observed to be much higher than those of other MnO₂ adsorbents, which resulted from its largest BET surface area. Manganese oxides with a larger BET surface area possessed more site vacancies, resulting in more surface hydroxyl groups. These results suggested that the BET surface area was the predominant factor affecting the content of hydroxyl groups on manganese oxides.

Table 4 Parameters for modeling the acid–base surface chemistry of MnO₂ adsorbents

Sorbents	SBET (m^2 g^−1)	≡XOH (mmol g^−1)	≡YOH (mmol g^−1)	log K≡XO^−	log K≡YO^−	log K≡YOH2^+
α-MnO2	97.5	0.1996	0.0355	−0.86	−1.45	4.35
β-MnO2	14.2	0.0477	0.0333	−1.24	−1.91	4.61
γ-MnO2	71.4	0.1993	0.0214	−1.25	−2.67	4.71
δ-MnO2	270.8	0.7093	0.4090	−1.76	−4.52	2.57
λ-MnO2	36.6	0.1528	0.0208	−1.44	−1.28	4.28

---

A two-site model is employed to describe the reactions at manganese oxides surface, and the following reactions were used in FITEQL 3.2 to model Pb(II) adsorption on the five MnO₂ adsorbents surface.

On the DCS sites:

≡XOH + H⁺ = ≡XOH₂⁺ (11)

≡XOH = ≡XO⁻ + H⁺ (12)

≡XOH₂⁺ + Pb²⁺ = ≡XOPb⁺ + 2H⁺ (13)

≡XO⁻ + Pb²⁺ + H₂O = ≡XOPbOH + H⁺ (14)

On the DES sites:

≡YOH + H⁺ = ≡YOH₂⁺ (15)

≡YOH = ≡YO⁻ + H⁺ (16)

≡YOH²⁺ + Pb²⁺ = ≡YOPb⁺ + 2H⁺ (17)

≡YO⁻ + Pb²⁺ + H₂O = ≡YOPbOH + H⁺ (18)

The FITEQL failed to simulate the surface species on the five MnO₂ adsorbents because the species ≡XOH₂⁺ did not exist below the IEP of MnO₂.⁴² The zeta potential analysis showed that the IEP values of the five MnO₂ adsorbents were all below tested solution pH, suggesting that the ≡XOH₂⁺ and ≡XOPb⁺ would not be significant species in the present study. Thus, the reaction ≡XOH₂⁺ + Pb²⁺ = ≡XOPb⁺ + 2H⁺ was ignored in this study. All the parameters used to simulate the acid–base chemistry of the five MnO₂ adsorbents are summarized in Table 4.

Fig. 9 presents predicted the surface binding species for Pb(II) adsorbed onto the five MnO₂ adsorbents in 0.001, 0.01 and 0.1 mol L⁻¹ NaNO₃. It is apparent that the fitted curves provided a close description of Pb(II) adsorption on the five MnO₂ adsorbents over the tested pH range. Within the pH range of 2.5–10.0, ≡XOPbOH was the predominant species on α-, β-, γ- and λ-MnO₂ surfaces, while ≡YOPb⁺ made insignificant contributions to adsorption at acidic to neutral pH conditions. The ≡YOPb⁺ species decreased with the increase of solution pH, probably because the OH⁻ species increased at higher pH conditions and would bind to the ≡YOPb⁺, resulting in the increase of ≡YOPbOH species on α-, β-, γ- and λ-MnO₂ surfaces. For δ-MnO₂, ≡YOPb⁺ was the predominant species over the pH range of 2.5–5.8, likely due to the large amount of ≡YOH groups on the layered structural surfaces. Notably, Pb(OH)₂ species appeared on the surface of β- and λ-MnO₂ at high pH conditions, suggesting the Pb(II) had precipitated. This is conceivably attributed to the insufficient hydroxyl groups on β- and λ-MnO₂ surfaces for Pb(II) adsorption.

The adsorption isotherm data were reproduced by using parameters estimated from SCM, and the results are shown in Fig. 11. According to the fitting results from SCM, Pb(OH)⁺ and Pb(OH)₂ precipitate did not present at low pH conditions (<6.0) (Fig. 9). The aqueous Pb(II) species were Pb²⁺ and Pb(NO₃)⁺. For α-, β-, γ- and λ-MnO₂, ≡XOPbOH was the dominant species contributing to Pb(II) adsorption at the tested Pb(II) concentration in deionized water. This is likely due to the high Pb(II) concentration in the adsorption isotherm experiment. Similar results were also reported by Usiyama and Fukushi.⁴¹ As for δ-MnO₂, prediction of surface speciation showed that the ≡YOPb⁺ was the dominant species at various Pb(II) concentrations. The difference of dominant surface speciation between δ-MnO₂ and other MnO₂ adsorbents was possibly attributed to the high content of DES adsorption sites and high cation exchange capacity on layered manganese oxides.

Fig. 11 Modeling Pb(II) adsorption isotherms using diffuse double layer surface complexation model (SCM).

3.6 Fixed-bed adsorption

Fig. 12a illustrated an effluent history of δ-MnO₂ fixed-bed column adsorption experiment, which was carried out to evaluate the sorbent applicability in wastewater treatment. The breakthrough point was set as 0.5 mg L⁻¹ for Pb(II), as the regulated maximum concentration level by the mining industry emission standard of China.⁴⁶ The sorbent exhibited a breakthrough at the point of ∼25 000 bed volume (BV) under the experimental conditions, indicated that Pb(II) could be efficiently removed by δ-MnO₂ till the significant breakthrough occurs. The column adsorption behavior of Pb(II) on δ-MnO₂ could be further quantified by the Thomas model and Yoon–Nelson model. The Thomas model can be expressed as follows:⁴⁷

Fig. 12 Fixed-bed column adsorption and regeneration test. (a) Pb(II) removal from a synthetic wastewater by a column bed packed with δ-MnO₂. Influent Pb(II) was 20 mg L⁻¹ with 100 mg L⁻¹ of Na(I), 100 mg L⁻¹ Mg(II), and 150 mg L⁻¹ Ca(II), and the flow rate was 15.0 mL h⁻¹. (b) Regeneration of Pb(II) preloaded δ-MnO₂ using a mix solution containing 0.1 mol L⁻¹ HNO₃ and 5% (wt) Ca(NO₃)₂, and the flow rate was 7.5 mL h⁻¹.

The Yoon–Nelson model is given in eqn (20).⁴⁸

where C₀ is influent Pb(II) concentration (mg L⁻¹), C is effluent Pb(II) concentration (mg L⁻¹) at time t (h), k_T (L mg⁻¹ h⁻¹) and k_YN (min⁻¹) are the kinetic constants of Thomas model and Yoon–Nelson model, respectively, Q_m is the maximum solid-phase concentration of the solute (mg g⁻¹), M is the sorbents packed in the column (g), F is the flow rate (mL h⁻¹), and τ is the time required for 50% adsorbate breakthrough (h).

The fitting parameters of experimental data are summarized in Table 5. From the regression coefficients it could be concluded that the Thomas and Yoon–Nelson models fitted the experimental data accurately. As was not observed in the present study, the column breakthrough capacity was estimated using the Thomas model according to the literature.⁴⁸ The Pb(II) adsorption capacity at the breakthrough point was 6.27 × 10⁵ mg g⁻¹, which was much higher than the values reported for hydrous Zr(IV) oxide-based nanocomposite (319.4 mg g⁻¹) and polyacrylamide–hydrated ferric oxide hybrid material (211.4 mg g⁻¹).^48,49 Considering the structure of δ-MnO₂, the higher adsorption capacity was likely due to the capillary condensation, which was used advantageously as universal functionalization strategy in nanoparticle thin films assembled layer-by-layer.⁴⁹ The value of τ was as large as 7.32 × 10³ h, also remarkably larger than the reported values.^50,51 These results indicate that δ-MnO₂ is highly effective for removal of Pb(II) from water in both batch and fixed-bed mode. Notably, the synthetic method of δ-MnO₂ in this study was simple, economic, environmentally friendly, and was suitable for large-scale production. Therefore, δ-MnO₂ could be applied as a cost-effective adsorbent for the adsorptive removal of Pb(II) from wastewater.

Table 5 Estimated parameters of Thomas and Yoon–Nelson models for the sorption of Pb(II) by the δ-MnO₂ packed column

C0 (mg L^−1)	Flow rate (mL h^−1)	Thomas model			Yoon–Nelson model
		kT (L mg^−1 h^−1)	Qm (mg g^−1)	R^2	kYN (min^−1)	τ	R^2
20.0	15.0	4.26 × 10^−5	6.27 × 10^5	0.82	8.52 × 10^−4	7.32 × 10^3	0.82

---

The suppressed adsorption of Pb(II) on δ-MnO₂ observed at low pH implied that acidic treatment was a feasible approach to regenerating the Pb(II)-loaded adsorbent. The mixed regenerant of 0.1 mol L⁻¹ HNO₃ and 5 wt% wt Ca(NO₃)₂ was used for the exhausted δ-MnO₂ column regeneration at room temperature, and the results are showed in Fig. 12b. The cumulative desorption efficiency was around 20%, and about 80% of Pb(II) preloaded by δ-MnO₂ could not be effectively desorbed by the regeneration solution. Such low desorption ratio suggested that the adsorption sites of δ-MnO₂ were irreversible and could not be reused. Li et al.³⁵ and Pan et al.⁵² investigated the adsorption sites of δ-MnO₂ by EXAFS and found two distinct sites of δ-MnO₂ for heavy metal ions adsorption. They concluded that the edge-linkage inner-sphere complexation were irreversible, while the corner-linkage mode was reversible. Additionally, the ionic strengths effects proved surface complexes were formed by Pb(II) and surface adsorption sites. Thus, we suggested that the edge-linkage inner-sphere complexes were responsible for the irreversible Pb(II) adsorption by δ-MnO₂. Moreover, the ability of δ-MnO₂ fixed-bed column to efficiently remove Pb(II) from synthetic wastewaters makes it a good system for large-scale application.

4. Conclusions

In this work, we investigated the adsorption behavior of Pb(II) on five manganese oxides with different crystallographic phases. δ-MnO₂ showed the highest adsorption capacity due to its high BET surface area and layered structure which was fully accessible for Pb(II). Additionally, the large interlayer spacing and high amount of surface chemisorbed oxygen enhanced the Pb(II) adsorption on δ-MnO₂. On the contrary, adsorption capacities of porous α-, β-, γ- and λ-MnO₂ were constrained by the low efficiency of pores and Pb(II) was mainly adsorbed on the external surfaces. The adsorption kinetics of Pb(II) on MnO₂ adsorbents followed the pseudo-second-order model, and the adsorption of δ-MnO₂ was no affected by coexisting cations in water. The hydroxyl groups on the MnO₂ surface undergo pH dependent inner-sphere surface complexation reactions. The predominant species on α-, β-, γ- and λ-MnO₂ surfaces was ≡XOPbOH, while the ≡YOPb⁺ species was the predominant species on δ-MnO₂ surface over the pH range of 2.5–5.8. Column test verified the high working adsorption capacity and durability of δ-MnO₂ and therefore highlighted its promising potential as an effective adsorbent for the removal of heavy metal ions in water.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

Financial support from the Special S&T Project on Treatment and Control of Water Pollution of China (No. 2017ZX07201005) is gratefully acknowledged.

References

1. M. Liu, Y. Wang, L. Chen, Y. Zhang and Z. Lin, ACS Appl. Mater. Interfaces, 2015, 7, 7961–7969.
2. C. H. Yan, J. Xu and X. M. Shen, Environ. Health Perspect., 2013, 121, A294–A295.
3. S. Hou, L. Yuan, P. Jin, B. Ding, N. Qin, L. Li, X. Liu, Z. Wu, G. Zhao and Y. Deng, Theor. Biol. Med. Modell., 2013, 10, 13.
4. World Health Organization, Guidelines for Drinking-water Quality, Incorporating The First and Second Addenda, 3rd edn, 2008, vol. 1, http://www.who.int/water_sanitation_health/dwq/fulltext.pdf.
5. Z. X. Wu and D. Y. Zhao, Chem. Commun., 2011, 47, 3332–3338.
6. X. Gu, L. J. Evans and S. J. Barabash, Geochim. Cosmochim. Acta, 2010, 74, 5718–5728.
7. C. C. Wang, L. C. Juang, C. K. Lee, T. C. Hsu, J. F. Lee and H. P. Chao, J. Colloid Interface Sci., 2004, 280, 27–35.
8. D. Dong, Y. M. Nelson, L. W. Lion, M. L. Shuler and W. C. Ghiorse, Water Res., 2000, 34, 427–436.
9. L. D. Puppa, M. Komárek, F. Bordas, J. C. Bollinger and E. Joussein, J. Colloid Interface Sci., 2013, 399, 99–106.
10. Y. Wang, X. Feng, M. Villalobos, W. Tan and F. Liu, Chem. Geol., 2012, 292–293, 25–34.
11. S. M. Maliyekkal, K. P. Lisha and T. Pradeep, J. Hazard. Mater., 2010, 181, 986–995.
12. R. M. McKenzie, Soil Res., 1981, 19, 41–50.
13. J. W. Tonkin, L. S. Balistrieri and J. W. Murray, Appl. Geochem., 2004, 19, 29–53.
14. Y. C. Son, V. D. Makwana, A. R. Howell and S. L. Suib, Angew. Chem., Int. Ed., 2001, 40, 4280–4283.
15. M. Yano, S. Suzuki, M. Miyayama and M. Ohgaki, Solid State Ionics, 2013, 233, 32–37.
16. X. B. Luo, K. Zhang, J. M. Luo, S. L. Luo and J. Crittenden, Environ. Sci. Technol., 2016, 50, 13002–13012.
17. M. G. Kozlova, S. V. Balakhonov, E. A. Goodilin, B. R. Churaulov, A. G. Veresov and Y. D. Tret'yakov, Russ. Chem. Bull., 2008, 57, 1151–1156.
18. G. V. Korshin, H. S. Chang, A. I. Frenkel and J. F. Ferguson, Environ. Sci. Technol., 2007, 41, 2560–2565.
19. J. Al Abdullah, A. G. Al Lafi, W. Al Masri, Y. Amin and T. Alnama, Water, Air, Soil Pollut., 2016, 227, 1–14.
20. X. Wang and Y. D. Li, Chemistry, 2003, 9, 300–306.
21. H. P. Zhang, L. Q. Gu, L. Zhang, S. R. Zheng, H. Q. Wan, J. Y. Sun, D. Q. Zhu and Z. Y. Xu, Appl. Surf. Sci., 2017, 406, 330–338.
22. H. Chen, Y. Wang and Y. K. Lv, RSC Adv., 2016, 6, 54032.
23. H. M. Xu, Z. Qu, S. J. Zhao, J. Mei, F. Q. Quan and N. Q. Yan, J. Hazard. Mater., 2015, 299, 86–93.
24. Q. H. Zhang, S. Sun, S. Li, H. Jiang and J. G. Yu, Chem. Eng. Sci., 2007, 62, 4869–4874.
25. S. B. Kanungo and K. M. Parida, J. Colloid Interface Sci., 1984, 98, 245–251.
26. D. M. Robinson, Y. B. Go, M. Greenblatt and G. C. Dismukes, J. Am. Chem. Soc., 2010, 132, 11467–11469.
27. L. P. Kang, M. M. Zhang, Z. H. Liu and K. Ooi, Spectrochim. Acta, Part A, 2007, 67, 864–869.
28. W. A. Jefferson, C. Hu, H. Liu and J. Qu, Chemosphere, 2015, 119, 1–7.
29. H. W. Nesbitt, G. W. Canning and G. M. Bancroft, Geochim. Cosmochim. Acta, 1998, 62, 2097–2110.
30. S. M. Ponder, J. C. Darab and T. E. Mallouk, Environ. Sci. Technol., 2000, 34, 2564–2569.
31. Z. Rao, K. Feng, B. Tang and P. Wu, ACS Appl. Mater. Interfaces, 2017, 9, 2594–2605.
32. D. C. Golden, C. C. Chen and J. B. Dixon, Clays Clay Miner., 1987, 35, 271–278.
33. H. S. Kim and P. C. Stair, J. Phys. Chem. B, 2004, 108, 17019–17026.
34. K. D. Kwon, K. Refson and G. Sposito, Geochim. Cosmochim. Acta, 2010, 74, 6731–6740.
35. X. Li, G. Pan, Y. Qin, T. Hu, Z. Wu and Y. Xie, J. Colloid Interface Sci., 2004, 271, 35–40.
36. A. Perner, K. Holl, D. Ilic and M. Wohlfahrt-Mechrens, Eur. J. Inorg. Chem., 2002, 5, 1108–1114.
37. M. Sekar, V. Sakthi and S. Rengaraj, J. Colloid Interface Sci., 2004, 279, 307–313.
38. J. Wang, S. Zheng, Y. Shao, J. Liu, Z. Xu and D. Zhu, J. Colloid Interface Sci., 2010, 349, 293–299.
39. M. B. Jazi, M. Arshadi, M. J. Amiri and A. Gil, J. Colloid Interface Sci., 2014, 422, 16–24.
40. T. Usiyama and K. Fukushi, Geochim. Cosmochim. Acta, 2016, 190, 134–155.
41. A. M. Puziy, O. I. Poddubnaya, V. N. Zaitsev and O. P. Konoplitska, Appl. Surf. Sci., 2004, 221, 421–429.
42. Y. Al Degs, M. A. M. Khraisheh and M. F. Tutunji, Water Res., 2001, 35, 3724–3728.
43. R. Han, W. Zou, Z. Zhang, J. Shi and J. Yang, J. Hazard. Mater., 2006, 137, 384–395.
44. P. J. Pretorius and P. W. Linder, Appl. Geochem., 2001, 16, 1067–1082.
45. J. W. Tonkin, L. S. Balistrieri and J. W. Murray, Appl. Geochem., 2004, 19, 29–53.
46. Integrated Water Discharge Standard, Ministry of the People's Republic of China, Beijing, 1996.
47. H. C. Thomas, J. Am. Chem. Soc., 1944, 66, 1664–1666.
48. J. F. Costa, V. J. Vilar, C. M. Botelho, E. A. Da Silva and R. A. Boaventura, Water Res., 2010, 44, 3946–3958.
49. M. Hua, Y. N. Jiang, B. Wu, B. C. Pan, X. Zhao and Q. X. Zhang, ACS Appl. Mater. Interfaces, 2013, 5, 12135–12142.
50. F. Fu and Q. Wang, J. Environ. Manage., 2011, 92, 407–418.
51. H. L. Lu, W. H. Zhang, Y. X. Yang, X. F. Huang, S. Z. Wang and R. L. Qiu, Water Res., 2012, 46, 854–862.
52. G. Pan, Y. Qin, X. Li, T. Hu, Z. Wu and Y. Xie, J. Colloid Interface Sci., 2004, 271, 28–34.

---