Materials

Silica gel (SiO₂•0.002H₂O, silica gel 60, particle size of 0.063–0.100 mm, Merck KGaA, Darmstadt, Germany), sodium hydroxide pellets (NaOH >97%, Ajax Finechem, New South Wales, Australia), lead(II) acetate trihydrate ((CH₃COO)₂Pb·3H₂O, abbreviated as Pb(ac)₂, >99.5%, Merck KGaA, Darmstadt, Germany), and aqueous HNO₃ solution (65%, Merck KGaA, Darmstadt, Germany) were used without further purification. Milli-Q water (ELGA, model: OS007BPM1, Type II 15 MΩ•cm, High Wycombe, UK) was used throughout.

Preparation of Magadiite and Octosilicate

Na-magadiite was prepared by the hydrothermal reaction of NaOH and SiO₂ as reported previously (Kosuge et al., Reference Kosuge, Yamazaki, Tsunashima and Otsuka1992; Ide & Ogawa, Reference Ide and Ogawa2007). NaOH (4.14 g) was dissolved in water (150 mL) and 28.4 g of silica gel was added to the NaOH solution. The gel was stirred for 1 h at room temperature. The mixture was sealed in a Teflon-lined stainless steel bottle and treated hydrothermally at 150°C for 2 days. The product was collected by centrifugation, washed with a dilute aqueous solution of NaOH (pH 10.0), and dried in air at 40°C for 2 days. Na-octosilicate was prepared by a hydrothermal reaction as reported previously (Endo et al., Reference Endo, Sugahara and Kuroda1994). NaOH (25.8 g) was dissolved in water (150 mL) and 81.5 g of silica gel was added to the NaOH solution. The gel was stirred for 1 h at room temperature. The mixture was sealed in a Teflon-lined stainless steel bottle and treated hydrothermally at 100°C for 1 month. The product was collected by centrifugation and dried in air at 40°C for 2 days.

Adsorption of Pb ²⁺ from Aqueous Solution

Adsorption of Pb²⁺ onto the layered alkali silicates was examined by means of the reaction between the layered alkali silicates and aqueous solution of Pb(ac)₂. A layered alkali silicate (0.1 g) was dispersed in water (25 mL) and stirred magnetically for 1 h. An aqueous solution of Pb(ac)₂ (25 mL) was mixed with the aqueous suspension of the layered alkali silicates and the mixture was stirred magnetically at room temperature for 24 h. The initial concentration of Pb ²⁺ was 0.2, 0.6, 1.0, 2.2, 2.8, 3.6, and 4.0 mM for magadiite and 0.2, 0.6, 1.0, 3.7, 4.0, 5.0, and 6.0 mM for octosilicate. The solids were collected by centrifugation using models CR22N and R20A2 rotors, from Hitachi Koki (Tokyo, Japan) at 20,000 rpm (48,000×g) for 10 min and dried in air at 40°C for 2 days. The concentration of Pb²⁺ in the supernatant was determined by inductively coupled plasma-optical emission spectroscopy (ICP-OES), from Agilent Technologies 700 Series instrument (California, USA). Adsorption of Pb²⁺ onto the layered alkali silicates from acidic aqueous solution was also examined using the aqueous solution of the Pb(ac)₂ (3.7 mM and pH = 2, which was adjusted by the addition of HNO₃ solution).

The ion-exchange reactions were performed by varying reaction times (2, 5, 10, 30, 60, 120, and 180 min) in order to follow Pb²⁺ adsorption kinetics by the following condition: a layered alkali silicate (0.1 g) was dispersed in water (25 mL) and stirred magnetically for 1 h. Pb(ac)₂ solution (25 mL, 2.2 mM for magadiite and 3.7 mM for octosilicate) was added into the suspension. The solids were collected by centrifugation using models CR22N and R20A2 rotors, Hitachi Koki (Tokyo, Japan) at 20,000 rpm (48,000×g) for 10 min after mixing for various times (2, 5, 10, 30, 60, 120 and 180 min) at room temperature. The concentration of Pb²⁺ in the supernatant was determined by ICP-OES.

Characterization

X-ray powder diffraction (XRD) patterns were recorded on a Bruker (Karlsruhe, Germany) NEW D8 ADVANCE X-ray powder diffractometer using Ni-filtered CuKα irradiation operated at 40 kV and 40 mA. Inductively coupled plasma-optical emission spectroscopy (ICP-OES) was performed on an Agilent Technologies (California, USA) 700 Series instrument. Calibration curves (R factor > 0.99) were made for each measurement using standard Pb(ac)₂ solution. Scanning electron microscopy (SEM) was done using a JEOL (Tokyo, Japan) JSM-7610F scanning electron microscope for samples coated with Pt. Elemental mapping was obtained using an energy dispersive X-ray fluorescence spectrometer (X-Max 150 mm², Oxford, UK) attached to the SEM.

Preparation of Magadiite and Octosilicate

The formation of magadiite and octosilicate (Kosuge et al., Reference Kosuge, Yamazaki, Tsunashima and Otsuka1992; Endo et al., Reference Endo, Sugahara and Kuroda1994; Ide & Ogawa, Reference Ide and Ogawa2007) was confirmed by the XRD patterns (Figs. 1a and 2a) and SEM images (Figs. 3a and 4a).

Fig. 1 XRD patterns of magadiite: a before and b–h after the reactions with initial aqueous solutions of Pb(ac)₂ of b 0.2, c 0.6, d 1.0, e 2.2, f 2.8, g 3.6, and h 4.0 mM

Fig. 2 XRD patterns of octosilicate: a before and b–h after the reaction the with initial aqueous solutions of Pb(ac)₂ of b 0.2, c 0.6, d 1.0, e 3.7, f 4.0, g 5.0, and h 6.0 mM

Fig. 3 SEM images and elemental mapping of magadiite: a before and b after the reaction with a Pb(ac)₂ solution concentration of 4.0 mM

Fig. 4 SEM images and elemental mapping of octosilicate: a before and b after the reaction with a Pb(ac)₂ solution concentration of 6.0 mM

Adsorption of Pb ²⁺ on Magadiite and Octosilicate

The XRD patterns of magadiite and octosilicate after reaction with Pb²⁺ are shown in Figs. 1 and 2. The basal spacing (d ₀₀₁) of magadiite decreased from 1.57 nm to 1.42 nm after the reaction with Pb²⁺, suggesting ion exchange between Pb²⁺ and Na in the interlayer space. The intensity of the d ₀₀₁ reflection became weaker after the adsorption of Pb²⁺. The interlayer space was calculated from the observed basal spacing (1.42 nm) by subtracting the thickness of the silica layer of magadiite, which was estimated from the basal spacing of the H-magadiite (Rojo et al. Reference Rojo, Ruiz-Hitzky and Sanz1988), to be 0.30 nm, thus giving a final value of 1.12 nm. The observed basal spacing (1.42 nm) was similar to that (1.43 nm) reported for the Pb²⁺-exchanged magadiite (Hatsushika, Reference Hatsushika1996). The basal spacing (d ₀₀₁) of octosilicate decreased from 1.10 nm to 0.88 nm after reaction with Pb²⁺, suggesting ion exchange between Na and Pb²⁺ in the interlayer space. Because the thickness of the silicate layer of octosilicate is 0.74 nm (Vortmann et al., Reference Vortmann, Rius, Siegmann and Gies1997), the interlayer space of the Pb²⁺-adsorbed octosilicate was estimated (by subtracting the thickness of the silicate layer from the observed basal spacing) to be 0.14 nm. The observed basal spacing (0.88 nm) is consistent with the value (0.85 nm) reported for Pb²⁺ exchanged octosilicate (Lim et al., Reference Lim, Jang, Park, Paek, Kim and Park2017). Those authors reported that the shrinkage of the basal spacing was caused by dehydration after the adsorption of Pb²⁺. The difference in the interlayer spaces for Pb²⁺ adsorbed magadiite (0.30 nm) and octosilicate (0.14 nm) in the present study was due to the difference in the hydration of the interlayer Pb²⁺ as well as the silicate framework deterioration.

The SEM images and elemental mapping of magadiite (Fig. 3) and octosilicate (Fig. 4) before and after reaction with the aqueous solution of Pb(ac)₂ (the initial concentration of 4.0 mM for magadiite and 6.0 mM of octosilicate) revealed that, even though the XRD peaks became weaker after the reaction with Pb²⁺ (Figs. 1 and 2), the flower-shaped morphology (4–5 μm) of magadiite composed of aggregated platy particles (1–2 μm) and the platy morphology of octosilicate (2–4 μm) were retained. Pb²⁺ was distributed homogeneously in/on the silicate particles for both magadiite (Fig. 3) and octosilicate (Fig. 4) systems, confirming the successful ion exchange.

Adsorption Isotherm of Pb ²⁺ on Magadiite and Octosilicate

The adsorption isotherms of Pb²⁺ on magadiite and octosilicate (Fig. 5) were of type H, according to the classification by Giles et al. (Reference Giles, MacEwan, Nakhwa and Smith1960), suggesting significant affinity between Pb²⁺ and the layered alkali silicates. The ideal cation exchange capacities of magadiite (2.2 meq/g) and octosilicate (3.7 meq/g), which are derived from the chemical formulae (magadiite, Na₂Si₁₄O₂₉·10H₂O, and octosilicate, Na₂Si₈O₁₇·8H₂O), are shown by the dotted lines (Fig. 5) to confirm that the amounts of Pb²⁺ adsorbed were larger than the calculated CEC values, especially for octosilicate. The adsorption of Pb²⁺ on magadiite reached the theoretical CEC value (2.2 meq/g) when the initial concentration of Pb(ac)₂ was 2.2 mM. When solutions of higher concentrations (2.8, 3.6, 4.0 mM) were used, the amounts of Pb²⁺ adsorbed were slightly more than the theoretical CEC value (2.2 meq/g) and the maximum amount of Pb²⁺ adsorbed was 1.27 mmol Pb/g magadiite (initial pH = 6.0). The adsorption of Pb²⁺ on octosilicate reached the theoretical CEC value (3.7 meq/g) when the initial concentration of Pb(ac)₂ was 3.7 mM. When solutions of greater concentrations (4.0, 5.0, 6.0 mM) were used, the amounts of Pb²⁺ adsorbed were greater than the theoretical CEC (3.7 meq/g) and the maximum was 2.34 mmol Pb/g octosilicate (initial pH = 6.0).

Fig. 5 Adsorption isotherms of Pb²⁺ on magadiite (triangle) and octosilicate (circle) from an aqueous solution at room temperature

The reported collection of Pb²⁺ from aquous solution by means of ion exchangers such as Na-bentonite (Glatstein & Francisca, Reference Glatstein and Francisca2015), carbon nanotubes (Li et al., Reference Li, Wang, Wei, Zhang, Xu, Luan, Wu and Wei2002; Abbas et al., Reference Abbas, Al-Amer, Laoui, Al-Marri, Nasser, Khraisheh and Atieh2016; Aliyu et al., Reference Aliyu, Kariim and Abdulkareem2017), activated carbon (Goel et al., Reference Goel, Kadirvelu, Rajagopal and Garg2005; Chen et al., Reference Chen, Pan, Chen, Yu, Wang and Yan2014), fly ash (Al-Zboon et al., Reference Al-Zboon, Al-Harahsheh and Hani2011; Barbosa et al., Reference Barbosa, Lapa, Lopes, Gunther, Dias and Mendes2014), zeolite (Jamil et al., Reference Jamil, Ibrahim, Abd El-Maksoud and El-Wakeel2010), and natural apatite (Kaludjerovic-Radoicic & Raicevic, Reference Kaludjerovic-Radoicic and Raicevic2010) is summarized in Table 1. The amounts of Pb²⁺ adsorbed by magadiite (1.27 mmol/g) and octosilicate (2.34 mmol/g) in the present study were greater than those reported for other adsorbents (Table 1), confirming the advantages of using magadiite and, in particular, octosilicate, as very high-capacity adsorbents of Pb²⁺.

Table 1 Examples of the amounts of Pb²⁺ adsorbed on layered alkali silicates and various other ion exchangers

Literature reports of the amount of Pb²⁺ adsorbed by magadiite vary widely (Table 1). Benkhatou et al. (Reference Benkhatou, Djelad, Sassi, Bouchekara and Bengueddach2016) reported 0.048 mmol Pb/g (initial pH = 7.0), whereas Lim et al. (Reference Lim, Jang, Park, Paek, Kim and Park2017) reported that the adsorption of Pb²⁺ from aqueous NaCl solution could reach as high as 0.97 mmol Pb/g magadiite and 1.5 mmol Pb/g octosilicate. The experimental conditions (pH, Pb²⁺ concentration, presence of competing ions, as well as the amount of adsorbent) may account for this difference. Because the negative charge of the layered alkali silicates is pH dependent (Schwieger et al., Reference Schwieger, Lagaly, Auerbach, Carrado and Dutta2004), protons may compete with Pb²⁺ for the exchange with Na at lower pH. For example, in the present study, the initial pH of the Pb(ac)₂ solution was in the range 5.9–6.0 and that of the silicate suspension was in the range 9.5–10.0. After mixing the solution and the suspension, the pH of the mixture was in the range 6.1–8.8 for the magadiite system and 6.6–9.3 for the octosilicate system. After reaction for 24 h, the ranges became 5.6–9.0 for magadiite and 5.8–9.1 for octosilicate, so the proton was not exchanged with Na. The adsorption of Pb²⁺ under these conditions was 1.27 mmol/g for magadiite and 2.34 mmol/g for octosilicate (Table 1), but from acidic solution, the amounts adsorbed decreased to 0.32 (pH 3.64) and 0.34 (pH 3.95), respectively, confirming the important role of pH.

The Langmuir model (Langmuir, Reference Langmuir1918) was used to describe the adsorption isotherms, and the parameters are summarized in Table 2, together with the reported values. The Langmuir model is expressed as:

Table 2 Langmuir parameters for the adsorption of Pb²⁺ on the layered alkali silicates and various other ion exchangers

$\frac{C_e}{q_e}=\frac{C_e}{q_m}+\frac{1}{q_m{K}_L}$

where C _e is the equilibrium concentration (mmol/L), q _e is the amount adsorbed (mmol/g), q _m is the maximum adsorption capacity of the adsorbent (mmol/g), and K_L is the Langmuir constant (L/mmol) related to the free energy of the adsorption (ΔG). By plotting C _e/q _e (vertical axis) versus C _e (horizontal axis), the values of q _m and K_L were determined from the slope and intercept of the linear plot, respectively. The Langmuir plot (Fig. 6a) fitted magadiite with a high correlation coefficient (R² = 0.9991). The Langmuir constant (K_L), which may correlate with the ΔG value (the larger K_L value corresponded to the more negative ΔG), was high (812 L/mmol) for magadiite when compared with the previous studes (Benkhatou et al., Reference Benkhatou, Djelad, Sassi, Bouchekara and Bengueddach2016; Lim et al., Reference Lim, Jang, Park, Paek, Kim and Park2017; Attar et al., Reference Attar, Demey, Bouazza and Sastre2019), suggesting the high affinity between the silicate surface and Pb²⁺ ions.

Fig. 6 Langmuir plots of the adsorption of Pb²⁺ onto a magadiite (triangles) and b octosilicate (circles)

In the case of octosilicate, the Langmuir model was fitted with a high correlation coefficient (R² = 0.9972) at concentrations between 0.6 and 6.0 mM (Fig. 6b). The deviation of the plot from the Langmuir model at concentrations of <0.6 mM has not been explained clearly thus far; the precipitation of Pb salt may have contributed to the change in the concentration of Pb²⁺. In order to discuss the point, the precipitation of Pb²⁺ salts was proposed as a mechanism for collection of Pb²⁺, in addition to the cation exchange. Precipitation of Pb salts may occur as a result of the change in pH caused by ion exchange. The Na released from the layered silicates caused an increase in pH as mentioned above; the pH of the Pb(ac)₂ solution was in the range 5.9–6.0, while that of the suspension after the reactions between the Pb²⁺ and the layered silicates was in the range 5.6–9.0 for magadiite and 5.8–9.1 for octosilicate. The precipitation of Pb₂(OH)Cl₃ during the reaction between magadiite and Pb(II) nitrate in aqueous NaCl solution was reported by Lim et al. (Reference Lim, Jang, Park, Paek, Kim and Park2017). The precipitation of In(OH)₃ during the reaction between layered alkali titanates and In(III) chloride in water was reported by Sirinakorn et al. (Reference Sirinakorn, Bureekaew and Ogawa2018b) and the precipitation of cadmium carbonate (Cd(CO₃)₂) during the reaction between layered alkali titanates and Cd(II) acetate in water was reported by Sirinakorn et al. (Reference Sirinakorn, Bureekaew and Ogawa2019). Pb(OH)₂ precipiates from water at a pH of >6.5 (Baes & Mesmer, Reference Baes and Mesmer1976) and PbCO₃ precipitates from water at pH < 9.5 (Taylor & Lopata, Reference Taylor and Lopata1984). The solubility product constants (K_sp) of Pb(OH)₂ and PbCO₃ are 1.2×10^–15 and 7.4×10^–14, respectively. In the present study, the pH of the suspension after the adsorption experiment was 5.6–9.0 for magadiite and 5.8–9.1 for octosilicate, so that the precipitation of Pb(OH)₂ and/or PbCO₃ was a plausible mechanism for the amount of Pb²⁺ adsorbed above the value of the CEC, especially in the case of octosilicate. The complicated Langmuir plot supported the hypothesis that the removal of Pb²⁺ by the reaction with octosilicate involved ion exchange and the precipitation of Pb salts.

Scanning electron microscopy images of magadiite and octosilicate before and after reaction with Pb(ac)₂ solution of 4.0 mM for magadiite and 6.0 mM for octosilicate are shown in Fig. 7. Irregularly shaped particles <100 nm wide were observed on the edge surface of the silicate particles (Fig. 7b,d). The particles were thought to be Pb(OH)₂ and/or PbCO₃, though crystalline Pb(OH)₂ and PbCO₃ were not detected in the XRD patterns (Figs 1, 2) and showing the absence of Si in the small particles by SEM/EDX analyses was not possible.

Fig. 7 SEM images of a magadiite and c octosilicate before and after the reaction with Pb(ac)₂ solution at a concentration of b 4.0 mM for magadiite and d 6.0 mM for octosilicate

Kinetics of Adsorption of Pb ²⁺ on Magadiite and Octosilicate

Pb²⁺ adsorption on magadiite reached equilibrium within 5 min, while the adsorption of Pb^{2 +} on octosilicate took 60 min to reach equilibrium. The reaction time reported for the adsorption of Pb²⁺ on magnetite (Hong et al., Reference Hong, Xie, Mirshahghassemi and Lead2020), hydroxylapatite (Dong et al., Reference Dong, Zhu, Qiu and Zhao2010; Googerdchian et al., Reference Googerdchian, Moheb and Emadi2012; Choudhury et al., Reference Choudhury, Mondal and Majumdar2015), ferroxane (Moattari et al., Reference Moattari, Rahimi, Rajabi, Derakhshan and Keyhani2015), attapulgite (Deng et al., Reference Deng, Gao, Liu, Hu, Wei and Sun2013), and carbon nanotubes (Wang et al., Reference Wang, Zhou, Peng, Yu and Yang2007; Kabbashi et al., Reference Kabbashi, Atieh, Al-Mamun, Mirghami, Alam and Yahya2009; Robati, Reference Robati2013) are summarized in Table 3. The short reaction time (5 min) for the removal of Pb²⁺ is an advantage of magadiite as an adsorbent. The particle morphology (platy particles a few μm wide for octosilicate and flower-like aggregates of platy particles 1–2 μm across for magadiite) affected the reaction time; the flower-like aggregates of magadiite particles were suitable for smooth diffusion of Pb²⁺. This characteristic is worthy of further investigation for the adsorption of various metal ions on magadiite.

Table 3 Examples of the kinetics of adsorption of Pb²⁺ on various ion exchangers