Clay minerals are effective adsorbents used for the remediation of toxic heavy metals from wastewater due to their large surface areas and great cation-exchange capacities. In this study, the removal of lead ions from aqueous solutions via adsorption was investigated using raw and iron-modified Turkish sepiolite. The aim of this study was to examine the effects of modification and environmental conditions on the sorptive properties of sepiolite samples. Initially, the raw sepiolite (Sep) and magnetic sepiolite/Fe2O3 composite (MagSep) prepared using the co-precipitation method were characterized via mineralogical and petrographical means and the physicochemical properties were determined. Then, the batch adsorption of lead (Pb2+) ions on the sepiolite samples was examined under various conditions (solution pH, adsorbent dosage, contact time, initial Pb2+ ion concentration, temperature, shaking rate). The adsorption capacity of MagSep was found to be greater than that of Sep under all experimental conditions. The results showed that the adsorption process followed a pseudo-second-order kinetic model, and the Langmuir isotherm best correlated with the experimental data. The maximum adsorption capacities were found to be 60.6 and 90.1 mg g–1 for Sep and MagSep, respectively. The characterization of the Pb-adsorbed sepiolite samples showed that lead formed covalent bonds with the sepiolite samples and attached to the sepiolite surface mainly through ion exchange. MagSep can be used efficiently in the field of wastewater treatment for the removal of Pb2+ ions as it does not release any toxic pollutants and can be separated easily with the use of a magnetic field.

Heavy metals are highly toxic pollutants, with lead being one of the most frequently encountered. The main sources of lead are metal platings, fertilizers and pesticides, ore smelting, factory chimneys, vehicle exhaust fumes, urban waste and lead-based battery and paint products (Jaishankar et al., 2014; Kumar et al., 2022; Raj & Das, 2023). Environmental pollution occurs as a consequence of lead accumulation in air, water and soil. Living organisms are exposed to heavy metals through inhalation, ingestion or skin penetration. The damage to human health due to heavy metal exposure has been recognized previously (Deng et al., 2022; Mitra et al., 2022). When lead gets into the body, it generates reactive oxygen species, which lead to oxidative stress, causing cellular damage (Collin et al., 2022). Lead interrupts physiological functions, encourages many respiratory syndromes and results in clinical disorders affecting the nervous, kidney, cardiovascular and other systems (Raj & Das, 2023). Hence, the levels of heavy metals in drinking water are limited by regulations and standards (Uddin, 2017; Tamjidi et al., 2019); the level of lead is limited to 10 μg L–1 as stated by the World Health Organization (WHO) and the European Union (WHO, 2017; EU, 2020; Dettori et al., 2022). This indicates the importance of the remediation of heavy metals from the environment.

Heavy metals have typically been removed from wastewater using the redox process, reverse osmosis, electrochemical treatment, bioremediation, coagulation, precipitation, ion exchange, membrane filtration and adsorption depending on the available budget and the initial heavy metal concentration (Rajendran et al., 2022). Adsorption has become a popular method as its efficiency of removal can be improved with the development of original adsorbents and its cost is low compared to other separation processes (Fei & Hu, 2023). In adsorption, heavy metal ions in the aqueous solution attach successively to the surface of the adsorbent by means of physical or chemical interactions. The selection of the most efficient, easily accessible, low-cost, recyclable and environmentally friendly adsorbent is important to allaying environmental concerns (Novikau & Lujaniene, 2022). The ideal adsorbent should have a great accessible surface area (via a porous structure or small particle size), strong interaction between the active sites of the adsorbent and the heavy metals (employing surface modifications and/or optimum operating conditions), selectivity towards the targeted heavy metal species and easy regeneration (Fei & Hu, 2022).

Adsorbents are classified as carbon materials (activated carbon, carbon nanotubes, graphene and biochar), polymers (natural polymers such as chitosan and alginate, synthetic porous polymers such as amorphous porous organic polymers and crystalline covalent organic frameworks), metallic and metal compounds (nanoparticles, structurally engineered compounds and magnetic compounds) and others (minerals, boron and carbon nitrides, agricultural and industrial wastes and functionalized mesoporous adsorbents) according to their chemical components and material structures (Fei & Hu, 2022). From all of these, clay minerals are strong candidates for adsorption applications because of their global abundance, low cost and non-toxicity, in addition to their possession of a great surface area that enables the capture of cations and anions (Otunola & Olodale, 2020). For example, the use of bentonite, zeolite and perlite (Uygun et al., 2023), kaolinite, montmorillonite, goethite and ferrihydrite (Mao et al., 2023) and natural silicate minerals (Al-Degs et al., 2003) for the adsorption of lead(II) has been reported previously. It should be noted that during the use of alumina-silicate minerals in wastewater treatments, these minerals may release Al into the solution (Uygun et al., 2023).

Sepiolite, which has the chemical formula of Si12Mg8O30(OH)4(OH2)4⋅8H2O, is a natural, hydrated magnesium silicate clay mineral. It consists of a fibrous-like structure with fine microporous channels of various dimensions (Lazarević et al., 2007; Dogan et al., 2008). The structure of sepiolite is built up of ribbons of 2:1 layers, with each ribbon having a magnesium octahedral sheet between two layers of silica tetrahedra (Madejová et al., 2017), which extend as a continuous layer with an inversion of the apical ends every six units. The ribbons are linked by the inversion of the SiO4 tetrahedra through Si–O–Si bonds. This inversion produces a discontinuous octahedral sheet, in contrast to other clay minerals that contain continuous octahedral sheets. This gives sepiolite a structure with tunnel-like pores parallel to the fibre axis (Tartaglione et al., 2008). Exchangeable cations of magnesium are located at the edges of the octahedral sheets, completing the coordinate number with two structural water molecules (bound water). This bound water is also connected to the zeolitic water by hydrogen bonds that fill the spaces in the channels (Mahmoud et al., 2017). Previous studies have shown that sepiolite can remove toxic pollutants from the environment and retain a significant amount of heavy metal ions from aqueous solutions (Lazarević et al., 2007). Brigatti et al. (1996) stated that heavy metal cation sorption can occur at the surface of broken edges, in channels and at specific sites of sepiolite according to crystal–chemical affinity. The adsorption of NH3, H2S and SO2 (Zhou et al., 2023), ethylene gas (Erdoğan & Esenli, 2021), phenanthrene (González-Santamaría et al., 2017), Ni2+ ions (Kıpçak et al., 2020) and red dye (Çoruh et al., 2011) on sepiolite have also been reported previously.

The adsorption capacity of raw minerals can be improved through modifications, such as thermal, acid, organic and nano-zero-valent iron treatments, which increase their cation-exchange capacity (CEC) and surface area, resulting in increased adsorption capacity (Otunola & Olodale, 2020). The adsorption of various ions on sepiolite after acid treatment (Lazarević et al., 2007), iron impregnation (Yu et al., 2016; Fayazi et al., 2019) and chemical modification (Dogan et al.2008; Liang et al.2013) has been studied previously in the literature. With the incorporation of iron, the minerals gain magnetic properties, and the combination of adsorption and magnetic properties makes adsorbents potentially applicable in wastewater treatment due to their ease of separation (Tamjidi et al., 2019). The magnetic modification of chitosan (Wang et al., 2023), biogas slurry solids (Sasidharan & Kumar, 2022), activated carbon (González Vázquez et al., 2016) and mesoporous secondary nanostructures (Bhattacharya et al., 2015) has also been reported previously.

Türkiye has 13.5 million tonnes of sepiolite reserves (OIK, 2018), and 4.53% (612,000 tonnes) of this reserve is in the Eskişehir region. A few studies have reported previously on the adsorption of Ni2+ (Kıpçak et al., 2020), Cu2+ (Doğan et al., 2009), Mn2+, Cr3+ and Cd2+ (Kocaoba, 2009), Pb2+ (Bektaş et al., 2004) and Co2+ (Kara et al., 2003) using Turkish sepiolite. However, no study in the literature has reported the adsorption of Pb2+ using a magnetic sepiolite composite prepared from the sepiolite supplied from the Eskişehir region of Türkiye. Therefore, the aims of this study were: (1) to modify the surface of raw sepiolite supplied from Eskişehir, Türkiye, with iron; (2) to characterize both raw and modified sepiolite samples in detail; (3) to study the adsorption of Pb2+ ions from the aqueous solution using these sepiolite samples to observe the effects of surface modification on Pb2+ adsorption; (4) to fit the experimental data using various kinetic and adsorption isotherm models; (5) to discuss the adsorption mechanism of Pb2+ ions on the adsorbents; and (6) to characterize the adsorbents after Pb2+ ion adsorption.

Materials and reagents

Raw sepiolite (Sep), supplied from Eskişehir, Türkiye, was first treated to remove the water-soluble impurities and organic matter and then sieved to obtain a particle size of <63 μm. Pb(NO3)2 (≥99.0% purity, Sigma-Aldrich), Fe(NO3)3⋅9H2O (≥98% purity, Sigma-Aldrich) and all of the other chemicals used in this study were purchased as analytical grade.

Preparation of the magnetic sepiolite/Fe2O3 composite

The magnetic sepiolite/Fe2O3 composite (MagSep) was prepared using a procedure similar to that reported previously by Huang et al. (2017). Raw Sep was washed with Milli-Q® water and dried at 100°C in an oven to remove any impurities. To obtain MagSep, 7.05 g dry sepiolite was added to 300 mL Milli-Q water in a beaker, and the beaker was placed in an ultrasonic bath at room temperature. After dispersing the sepiolite fibres in water for 30 min, 16.16 g Fe(NO3)3⋅9H2O was added to the water-bearing sepiolite and left in the ultrasonic bath for 1 h. To co-precipitate iron and the sepiolite, 10 M NaOH was droppered into the beaker until the pH equalled 10.0, and this was again kept in an ultrasonic bath for 1 h. The composite was then heated to 90°C and left in the ultrasonic bath for an additional 2 h. It was then left for 12 h at room temperature and washed with Milli-Q water until the pH equalled 7.0, and this was then centrifuged. Later, the composite was washed with ethanol and filtered. Finally, the MagSep was dried in an oven at 80°C for 4 h. The preparation process of MagSep is shown in Fig. 1a.

Characterization of Sep and MagSep

Mineralogical and petrographic analyses. X-ray diffraction (XRD) analysis of the sepiolite samples was performed using an Inel Equinox 1000 XRD spectrometer equipped with a Co X-ray tube. XRD traces were recorded at 5–70°2θ at a step size of 0.02°. The elemental analysis of samples was performed using a PANalytical Epsilon Model 1 X-ray fluorescence (XRF) spectrometer instrument according to the TS EN 15309 standard. Images and the elemental composition of the samples were obtained using scanning electron microscopy (SEM)/energy-dispersive spectroscopy (EDS) with a ZEISS EVO 40 apparatus. The infrared spectra of the sepiolite samples were recorded in the 4000–400 cm–1 region on a Varian/660-IR Fourier-transform infrared (FTIR) spectrometer. The XRF, SEM-EDS and FTIR analyses of the adsorbents were conducted after the adsorption experiments.

Physicochemical properties of the sepiolite samples

The water-absorption capacity, CEC, dissolution in water and point of zero charge (PZC) values of the sepiolite samples were determined as explained in detail in a previous study (Uygun et al., 2023).

Specific surface area and Sauter mean particle diameter

The specific surface area and Sauter mean particle diameter of the sepiolite samples were determined using a Malvern Mastersizer 3000E device. The samples were dispersed in water and the measurements were performed at a stirring speed of 500 rpm.

Brunauer–Emmett–Teller surface area

The Brunauer–Emmett–Teller (BET) surface areas of the sepiolite samples were determined using a Quantachrome Autosorb 6B device.

ζ-potential

The ζ-potential values of the samples were determined at pH 7 using a Malvern Zetasizer Nano ZS90 device.

Vibrating-sample magnetometer analysis

The magnetic property of MagSep was determined using a Cryogenic Ltd PPMS device in vibrating-sample magnetometer (VSM) mode.

Batch adsorption experiments

The adsorption experiments of lead-contaminated waters were performed in 50 mL glass bottles according to ASTM 4646-03 (2004) to determine the effects of the initial pH (pHi) value of the solution (3.0–9.0), adsorbent dosage (1–10 g L–1), initial Pb2+ ion concentration (100–800 ppm), temperature (25–60°C), contact time (15–1440 min) and shaking rate (0–300 rpm). The effect of pH was examined by dispersing 0.25 g of adsorbent in a 50 mL Pb2+ solution at a concentration of 200 ppm. The pHi value of the solution was adjusted via the addition of 1.0 M HCl or 1.0 M NaOH solution. The effect of the shaking rate was determined using an orbital shaking incubator (Mipro MLI-120). After the batch adsorption experiments, the filtrate was separated from the adsorbent using filter paper and the Pb2+ concentration in the filtrate was determined using inductively coupled plasma optical emission spectroscopy (ICP-OES). A schematic diagram showing the batch adsorption experiments is given in Fig. 1b. Reproducibility experiments were also performed, and differences in Pb2+ ion concentrations under the same experimental conditions were ~5%.

The adsorption capacity (qt, mg g–1), removal rate (R, %) and distribution coefficient (Kd, mL g–1) were calculated using Equations 13, respectively:
$$q_t = ( {C_0-C_t} ) \displaystyle{V \over M}$$
1
$$R = \displaystyle{{C_0-C_t} \over {C_0}}100$$
2
$${\rm K}_{\rm d} = \displaystyle{{( C_0-C_t) {\rm \;}V} \over {C_tM}}$$
3
where C0 is the initial mass concentration of the adsorbate (mg L–1), Ct is the mass concentration of the adsorbate at any time t (mg L–1), V is the volume of the aqueous solution (L), M is the mass of the adsorbent (g) and qt = qe at Ct = Ce, with qe and Ce denoting the equilibrium adsorption capacity and equilibrium mass concentration, respectively.
Adsorption kinetics were investigated using three different models: the pseudo-first-order model, pseudo-second-order model and intra-particle diffusion model. Adsorption isotherms were determined using the Langmuir and Freundlich equations. The thermodynamic parameters such as Gibbs free energy of adsorption (ΔG°), heat of adsorption (ΔH°) and entropy change (ΔS°) related to the adsorption of Pb2+ ions on sepiolite samples were evaluated using Equations 4 and 5 and plotting ln(Kd) vs 1/T graphs:
$${\rm ln}( {\rm K}_{\rm d}) = \displaystyle{{{\rm \Delta }S^\circ } \over {\rm R}}-\displaystyle{{{\rm \Delta }H^\circ } \over {{\rm R}T}}$$
4
$${\rm \Delta }G^\circ{ \,=\, } {\rm \Delta }H-T{\rm \Delta }S = {-}{\rm R}T{\rm ln}( {{\rm K}_{\rm d}} ) $$
5
where T is temperature in Kelvin and R is the universal gas constant (8.314 J mol–1 K–1).

Mineralogical and petrographic characterization of the sepiolite samples

XRD spectra of sepiolite samples and Fe(NO3)3⋅9H2O were acquired to examine the crystal structure of the Sep and MagSep samples (Fig. 2). It can be seen that the XRD spectrum of Sep (Fig. 2a) has peaks belonging only to sepiolite, whereas MagSep (Fig. 2c) has peaks related to the addition of iron (Fig. 2b), which demonstrated the incorporation of iron into the crystal structure of the Sep.

Although the chemical composition of sepiolite may vary according to its origin, the main constituents persist as silicate, magnesium and calcium oxides. The elemental compositions of the sepiolite samples used in this study were determined using XRF (Table 1). The main components of Sep were Si and Mg, whereas the main components of MagSep were Si and Fe. The Si contents of Sep and MagSep were 29.89% and 20.72%, respectively, and the Mg contents were 16.62% and 0.40%, respectively. As can be seen from Table 1, Mg was replaced by Fe (24.07%) in MagSep. The removal of Mg cations from the edges of the octahedral layers with iron coating has also been reported previously (Eren & Gumus, 2011).

SEM images of the sepiolite samples are shown in Fig. 3. Sep showed a fibrous-like structure, with the fibres having a high length to diameter ratio, whereas this structure can be seen to be altered significantly in the MagSep images (Fig. 3). Although fibres can also be observed in the MagSep images, a more porous structure was obtained due to the structural change of the dispersed sepiolite fibres and the modification of the surface with iron(III) nitrate during MagSep synthesis. The results of the EDS analyses of samples are presented in Fig. 4 together with the elemental distributions and compositions of the samples as insets. The presence of iron can be determined from the 6.5 keV peak, which corresponds to 43.68% iron in the MagSep sample. In addition, the silica and magnesium contents of MagSep were lower than those of the Sep sample, which was also observed from the XRF data (Table 1).

The FTIR spectra of Sep, Fe(NO3)3⋅9H2O and MagSep were recorded in the 4000–400 cm–1 region (Fig. 5). In both the Sep and MagSep spectra, vibrations of the Mg–OH group were observed at 3690 cm–1 and coordinated water at 3566 cm–1 (Lazarević et al., 2010). The peaks at 1660 cm–1 (Sep) and 1654 cm–1 (MagSep) corresponded to OH stretching vibrations, representing zeolitic water in the channels (Doğan et al., 2008; Lazarević et al., 2010) and bound water coordinated to magnesium in the octahedral sheet (Doğan et al., 2008). The band at 1460 cm–1 was due to the hydroxyl bending vibration, which reflects the presence of bound water (Doğan et al., 2008). The bands at 1210, 1080 and 970 cm–1 in the FTIR spectra of Sep and MagSep represented the stretching of Si–O bonds (Doğan et al., 2008; Lazarević et al., 2010; Ahribesh et al., 2017), and the band at 1015 cm–1 exhibiting the Si–O–Si plane vibrations was due to the basal plane of the tetrahedral units (Tabak et al., 2009). There were bands at 690 and 640 cm–1, corresponding to the vibrations of the Mg–OH bond, and a band at 460 cm–1 originating from octahedral–tetrahedral bonds (Si–O–Mg bonds; Lazarević et al., 2010).

The band at 2394 cm–1 in the spectrum of iron nitrate appeared at 2350 cm–1 in the spectrum of MagSep, indicating a chemical interaction with iron occurring on the sepiolite surface during the modification. There was also a weak band at 1750 cm–1 in the FTIR spectrum of MagSep and iron nitrate, which could be due to the stretching of Fe–OH groups. In addition, the shift of the 1660 cm–1 band (Sep) to 1654 cm–1 (MagSep) and the 1460 cm–1 band (Sep) to 1416 cm–1 (MagSep) demonstrated the decrease in H2O content with the replacement of iron oxide molecules (Eren & Gumus, 2011). Thus, it can be stated that a successful surface modification of sepiolite was achieved and iron oxide was integrated into the sepiolite structure, forming covalent bonds between iron oxide and sepiolite.

Physicochemical properties of the sepiolite samples

The physicochemical properties of the sepiolite samples are presented in Table 2. The average particle size of Sep was 20.4 μm, which increased to 328.0 μm (MagSep) after the surface modification (Fig. 6). It can be seen that the number of particles with greater particle sizes increased after the modification. Thus, the specific surface area of Sep was reduced by 16 times in MagSep.

The great water-absorption capacities and low dissolution percentages make these sepiolite samples preferable materials for various areas of application such as wastewater remediation. From the results (Table 2), it can be said that the surface modification process decreased the water-absorption capacity of sepiolite by ~10%, and the dissolution percentage of sepiolite decreased by 51%, leading to less contamination of the treated water. The dissolved ions in the water were determined using ICP-OES analysis, and the data are presented together with the elemental analysis of Milli-Q water in Table 3 for comparison.

The ions that were most dissolved were Ca in Sep and Na in MagSep. Although Na was not present in great amounts in Sep (Table 1), the presence and dissolution of Na in water were due to the use of NaOH to attain a high pH while preparing the MagSep sample. It should also be mentioned that the ions released into the water are not toxic to the environment, which is an important concern relating to the pollution of remediated water sources. Iron is also included in Table 3 to demonstrate that iron was not dissolved after the surface modification as it was chemically bound to the surface. The metal cations occupy various positions inside the mineral, and their release depends on their structural positions. When they are weakly bonded inside the channels or on the mineral surfaces, they are partly but very rapidly released, whereas at the channel edges in the octahedra, they are bonded more strongly with similar exchange cations (Brigatti et al., 1996).

The BET surface area and CEC of the materials are the main physicochemical properties used for the selection of an adsorbent. The BET surface area of Sep was found to be 352.46 m2 g–1, which was comparable to the results obtained in studies performed using sepiolite supplied from Eskişehir, Türkiye (Doğan et al., 2008). The BET surface area of MagSep was found to be 9% lower than that of Sep (Table 2), which was due to the alteration of the fibrous-like structure of the sepiolite after surface modification by iron. The CEC of MagSep was doubled for the same reason. In the literature, the CEC of sepiolite has been reported to be between 10 and 45 meq 100 g–1 (Christidis, 2011).

VSM analysis was performed to determine whether the synthesis of MagSep was successful in terms of achieving magnetic properties. The magnetization curve of the MagSep sample is shown in Fig. 7. It can be seen that MagSep showed paramagnetic properties and the saturation value of magnetism was 0.5 emu g–1. The inset image in Fig. 7 demonstrates the behaviour of MagSep under an external magnetic field. It can be concluded that MagSep can easily be separated from aqueous solutions using permanent powerful magnets when used as an adsorbent.

Batch adsorption experiments

The effect of the pHi of the solution on Pb2+ adsorption on sepiolite samples. pHi is the most important parameter regarding adsorption as pH affects both the formation of lead compounds and the surface charge of the adsorbent. The effect of the pHi on the adsorption of Pb2+ ions in the contaminated aqueous solution by sepiolite samples was investigated in the pH range of 3.0–9.0 at 25°C for 1440 min, an initial Pb2+ ion concentration of 200 ppm, a shaking rate of 75 rpm and an adsorbent dosage of 5 g L–1.

The removal percentage of Pb2+ using Sep was lower compared to MagSep until pH = 7.0 and then equal at pH 7.0–9.0 (q = 39.9 mg g–1; Fig. 8). The adsorption capacity of the sepiolite modified by Fe3O4 and MnO2 was found to be equal to ~50 mg g–1 at pH 6-9 (Fayazi et al., 2019). The low removal percentages at low pH values (pH < PZC) can be attributed both to the positive charge density on the surface sites of sepiolite minerals that causes electrostatic repulsion between Pb2+ ions and the edge groups with positive charge (Si–OH2+) on the surface of sepiolite particles, and to the excess amount of H3O+ ions in the solution, which compete with the positively charged H+ ions and Pb2+ ions for the available adsorption sites on the sepiolite surface (Sharifipour et al., 2015). Hence, the pHi of the solution was taken as 7.0 in all of the batch adsorption experiments.

The ζ-potentials of Sep and MagSep, showing the potential difference between the dispersion medium and the stationary layer of the fluid of the sepiolite sample, at pH 7.0 were found to be –17.2 and –27.0 mV, respectively (Table 2). It can also be deduced from the Pourbaix diagram of lead that Pb(II) exists as a cation at pH 7.0 (Lehto & Hou, 2010). Thus, the removal of lead can be caused by the electrostatic attraction between the negatively charged sepiolite surface and the positively charged Pb2+ ions. In addition, the main mechanism for the adsorption of divalent cations on sepiolite powder is the ion exchange of Mg2+ ions from the sepiolite structure with metal ions (M2+) from the solution (Lazarević et al., 2007). This suggests that the retention of Pb2+ ions on the sepiolite sample surfaces took place through electrostatic interaction and exchange of Mg2+ ions together. The affinity for ion exchange being greatest in the case of Pb2+ ions was also reported by Lazarević et al. (2007).

The effect of contact time on Pb2+ adsorption on the sepiolite samples/adsorption kinetics

To investigate the effect of contact time on adsorption, experiments were carried out for 15–1440 min at pH 7.0, temperature 25°C, an initial Pb2+ ion concentration of 400 ppm, a shaking rate of 75 rpm and an adsorbent dosage of 5.0 g L–1. The removal percentages as a function of time are shown in Fig. 9a.

It was observed that the adsorption of lead ions on the sepiolite samples was completed at nearly 2 h for both samples. It can be said that the adsorption capacity increased by 2.35 mg g–1 (5.25%) with contact time when the adsorption capacity at 2 h was compared with 24 h. Hence, the remainder of the adsorption experiments were performed at 2 h. MagSep showed greater adsorption capacities at each studied contact time. Lead removal of 74.86% was achieved at 2 h, which was 21.72% greater than Sep, demonstrating the increase in adsorption performance with surface modification.

In a previous study, the adsorption of lead was studied under similar experimental conditions, and the adsorption capacities were in the following order: bentonite > zeolite > perlite (Uygun et al., 2023). When the removal percentages of Sep and MagSep at 2 h were compared with these minerals, the new order became: bentonite (95%) > MagSep (75%) > Sep (53%) > zeolite (42%) > perlite (27%). This demonstrates that bentonite has a greater affinity for lead ions.

Adsorption kinetics were studied to investigate the dependence on the physical and/or chemical characteristics of the sepiolite samples and thus the influence of contact time on the adsorption mechanism. The mechanism of adsorption was examined from the linear fitting of the kinetic data to the pseudo-first-order, pseudo-second-order and intra-particle diffusion models (Fig. 9b). The kinetic equations of the corresponding models are given in Table 4.

The rate constants (k1, k2, k3) were determined from the slopes, and the equilibrium adsorption capacities (qe) and intra-particle diffusion model constants (b) were determined from the intercepts of the linear equations (Fig. 9b & Table 4). It can be seen that lead adsorption was a time-dependent process and that the adsorption of Pb2+ on the sepiolite samples conformed with the second-order kinetic model (R2 = 0.9999 for Sep and 0.9995 for MagSep). The equilibrium adsorption capacities (qe) found from the model were also comparable (~3% difference) to the adsorption capacities found experimentally at 1440 min. Previous studies investigating the adsorption of lead on raw Sep (Bektaş et al., 2004; Sharifipour et al., 2015), Fe2O3–MnO2-modified sepiolite (Fayazi et al., 2019) and mercapto-functionalized sepiolite (Liang et al., 2013) have also reported that the experimental data conformed to pseudo-second-order kinetics.

The effect of adsorbent dosage on Pb2+ adsorption on the sepiolite samples

The effect of adsorbent dosage on the adsorption of Pb2+ from the contaminated aqueous solution by Sep and MagSep was studied by varying the adsorbent dosage between 1 and 10 g L–1 at pH 7.0, temperature 25°C, a contact time of 120 min, a shaking rate of 75 rpm and an initial Pb2+ ion concentration of 400 ppm (Fig. 10).

It is clear that the removal percentage of Pb2+ ions increased with the adsorbent dosage, as more Pb2+ ions were adsorbed from the greater amount of active sites on the surface of the sepiolite samples. The maximum adsorption capacities were 94.85 and 175.32 mg g–1 at the adsorbent dosage of 1 g L–1 for Sep and MagSep, respectively, whereas the greatest removal percentages were achieved at the adsorbent dosage of 10 g L–1, with values of 95.81% and 97.55% for Sep and MagSep, respectively. It was observed that the difference in adsorption capacities and, hence, the removal percentages between the two adsorbents decreased with increasing adsorbent dosages. The number of unsaturated adsorption sites on the adsorbent increased at high dosages, leading to a progressive decrease in adsorption capacities (Fayazi et al., 2019). Considering both the adsorption capacities and removal percentages of the sepiolite samples, an adsorbent dosage of 5 g L–1 was selected for the remainder of the adsorption experiments.

The effect of initial Pb2+ concentration on Pb2+ adsorption on the sepiolite samples/adsorption isotherms

The effect of initial Pb2+ concentration on the adsorption of Pb2+ on the sepiolite samples was studied by varying the initial Pb2+ ion concentrations between 100 and 800 ppm at pH 7.0, temperature 25°C, a contact time of 120 min, a shaking rate of 75 rpm and an adsorbent dosage of 5 g L–1 (Fig. 11a).

There was no difference between the adsorption capacities of Sep and MagSep at the initial adsorbate concentrations (100–200 ppm), and the removal percentages were >99.5% (Fig. 11a). Moreover, the difference between the adsorption capacities, and hence the removal percentages, increased with increasing adsorbate concentrations. As predicted, MagSep showed greater adsorption capacities than Sep at increased adsorbate concentrations, with the difference being 17.49% at 800 ppm. Similarly, the adsorption capacity of MagSep was 89 mg g–1, whereas it was 61 mg g–1 for Sep at the initial adsorbate concentration of 800 ppm. The increase in adsorption capacities with initial Pb2+ concentration is due to the fact that the mass transfer driving force would become larger at greater initial adsorbate concentrations (Fayazi et al., 2019). It can be concluded that aqueous solutions of 400 ppm can be adsorbed by MagSep with a removal percentage of 74.86%.

The Langmuir and the Freundlich equations are often used to describe adsorption equilibria for wastewater treatment applications. The adsorption isotherm that best fit the experimental data was determined using the linear forms of these models (Fig. 11b). The equations of the corresponding isotherm models are shown in Table 5.

The maximum adsorption capacities (qm) and Freundlich isotherm constants (n) were found from the slopes, and the Langmuir and Freundlich adsorption coefficients (KL, KF) were found from the intercepts of the corresponding linear equations (Fig. 11b & Table 5). By taking into account the correlation coefficients (R2), it can be seen that the Langmuir isotherm showed a better correlation over the studied concentration range for all adsorbents (R2 = 0.9718 for Sep and 0.9795 for MagSep). The maximum adsorption capacities (qm) were determined as 60.6 and 90.1 mg g–1 for Sep and MagSep, respectively. As the Langmuir isotherm fit the experimental data, this denotes that the sepiolite samples had a homogeneous distribution of active sites, indicating that monolayer adsorption on the surface of the sepiolite samples was observed (Huang et al., 2017). Previous studies investigating the adsorption of lead on raw and modified sepiolite have found that the data better fit the Langmuir isotherm (Bektaş et al., 2004; Fayazi et al., 2019) and Freundlich isotherm (Liang et al., 2013).

Studies investigating the adsorption of lead on the sepiolite minerals also found the maximum adsorption capacities to be 31.5 mg g–1 (Sharifipour et al., 2015), 51.36 mg g–1 (Eren & Gumus, 2011), 72.52 mg g–1 (Lazarević et al., 2007) and 93.54 mg g–1 (Bektaş et al., 2004) for raw Sep, 45.58 mg g–1 (Lazarević et al., 2007) for acid-activated sepiolite and 75.79 mg g–1 (Eren & Gumus, 2011) and 131.58 mg g–1 (Fayazi et al., 2019) for iron-modified sepiolite.

The effect of temperature on Pb2+ adsorption on the sepiolite samples/adsorption thermodynamics

The effect of temperature on the adsorption capacity of Pb2+ was studied at temperatures of 25–60°C at pH 7.0 for 120 min, an adsorbent dosage of 5 g L–1, a shaking rate of 75 rpm and an initial Pb2+ ion concentration of 400 ppm. It was found that the removal percentages increased with temperature, with lead removal rates of 64% (51.52 mg g–1) and 84% (67.23 mg g–1) at 60°C for Sep and MagSep, respectively. As the adsorption capacities of both sepiolite samples increased with temperature, the adsorption of Pb2+ on the sepiolite samples can be explained by chemical adsorption on the active sites of the samples. This was supported by the Langmuir adsorption isotherm determined previously, as in this model the ions are adsorbed on the adsorbate via covalent bonds, which are difficult to break at temperatures <100°C. To obtain high removal percentages of lead ions, temperatures >45°C and surface modification with iron impregnation of sepiolite can be suggested.

The enthalpy change (ΔH°) and entropy change (ΔS°) were obtained from the slope and intercept of the Van 't Hoff plot (Fig. 12) using Equation 4, and the Gibbs free energy of adsorption (ΔG°) was found from Equation 5.

The thermodynamic parameters of Sep and MagSep are listed in Table 6. Both Sep and MagSep showed a positive enthalpy change, demonstrating that the adsorption of lead ions on the sepiolite samples was endothermic. Furthermore, the observed enthalpy change was greater than that of typical ion-exchange reactions (typically <8.4 kJ mol–1; Liang et al., 2013), suggesting that specific interactions other than outer-sphere electrostatic interactions are operative in the Pb–sepiolite adsorption system. The negative values of the Gibbs free energy change for both sepiolite samples indicate that the adsorption process occurred spontaneously. The increasingly negative values of ΔG° with increasing temperature indicate that the adsorption reactions were more favourable thermodynamically at greater temperatures. Entropy change was positive for both adsorbents, reflecting an increase in randomness at the solid/liquid interface during the adsorption of Pb2+ on the sepiolite samples.

The enthalpy and entropy change of MagSep were greater than for Sep, while the reverse was true for Gibbs free energy change, leading to an increased adsorption capacity of MagSep compared to Sep (Table 6). Spontaneous and endothermic adsorption of lead on sepiolite samples has also been reported previously (Bektaş et al., 2004; Eren & Gumus, 2011; Liang et al., 2013; Fayazi et al., 2019).

The effect of shaking rate on Pb2+ adsorption on the sepiolite samples

The effect of shaking rate on Pb2+ ion adsorption was studied by varying the shaking rate between 0 and 300 rpm at pH 7.0, temperature 25°C, a contact time of 120 min, an adsorbent dosage of 5 g L–1 and an initial Pb2+ ion concentration of 400 ppm. It was found that the adsorption capacities of Sep and MagSep were comparable until 150 rpm, after which a steep increase was observed at 300 rpm for both of the adsorbents. The removal percentages of Sep and MagSep were 50.55 ± 4.14% (40.44 ± 3.31 mg g–1) and 70.93 ± 3.73% (56.40 ± 3.45 mg g–1) for 0–150 rpm, whereas they were 91.46% (73.17 mg g–1) and 94.88% (75.91 mg g–1) at 300 rpm. It can be concluded that the adsorption capacities can be increased to a greater extent at greater shaking rates as there was a more homogeneous distribution of the adsorbents in the aqueous solution, which also increased the chance of lead ions binding to the active sites of the sepiolite surfaces.

Characterization of the sepiolite samples after Pb2+ ion adsorption

After the batch adsorption experiment at pH 7.0, a contact time of 120 min, an adsorbent dosage of 5 g L–1, a shaking rate of 75 rpm and a temperature of 25°C using an initial Pb2+ ion concentration of 400 ppm, the adsorbed Pb2+ ions on the sepiolite samples were examined using XRF (Table 1), FTIR (Fig. 5) and SEM-EDS (Figs 3 & 4) analyses. It should be noted that the removal percentages of Pb2+ were found to be 53% and 75% for Sep and MagSep, respectively. The XRF analysis showed that Pb2+ was adsorbed successfully on the surfaces of Sep (3.12%) and MagSep (5.49%). The interaction between the surface of the sepiolite samples and Pb2+ ions can be argued to result from both electrostatic attraction due to the surface charge of the adsorbents and from ion exchange due to the exchange of cations, such as Mg and Ca. The decrease of Si after the adsorption (Table 1) also indicates the breakage of Si–O bonds during the adsorption process. The FTIR spectra (Fig. 5) show the changes in the positions and shapes of the fundamental vibrations of OH and Si–O groups due to the new Pb bonds formed. For example, the disappearing 1210 cm–1 band and the decreased intensity and the shift of the 1015 cm–1 band to 1001 cm–1 (Sep) and 1005 cm–1 (MagSep) indicate that Pb2+ ions influenced both Si–O bonds and Si–O–Si plane vibrations (Eren & Gumus, 2011; Hannachi et al., 2013; Kushwaha et al., 2019). The intense band at 2356 cm–1 in the spectra of Sep and MagSep indicates that a chemical interaction occurred on the sepiolite surface during the Pb2+ adsorption process. In addition, the shift in the 1660 cm–1 (Sep) and 1654 cm–1 (MagSep) bands before adsorption to 1647 cm–1 together with the disappearance of the 1460 cm–1 (Sep) and 1416 cm–1 (MagSep) bands after adsorption demonstrate the decrease of H2O content, being replaced by Pb2+ ions. The disappearance of the 690 cm–1 bands also demonstrate the interaction of Pb2+ with the Mg–OH bond (Eren & Gumus, 2011). Hence, all of these changes in the FTIR spectra are indicators of Pb2+ adsorption on the sepiolite samples, forming sepiolite–Pb covalent bonds.

The SEM images of the Pb-adsorbed sepiolite samples (Fig. 3) show that the surface structure of the samples was not altered by the adsorption process. The presence of Pb2+ on Sep (15.36%) and MagSep (21.89%) was also verified by the elemental analysis and distribution (Fig. 4).

The results of this study demonstrate the potential of Sep (Eskişehir, Türkiye) and MagSep as effective, environmentally friendly adsorbents for the removal of Pb2+ ions from aqueous solutions. The characterization of MagSep showed that iron was incorporated into the sepiolite structure through the formation of covalent bonds. The results also illustrated that MagSep had a greater adsorption capacity for Pb2+ ions compared to Sep under all experimental conditions due to the high affinity of iron oxide for lead ions. Therefore, MagSep can be termed an advanced adsorbent derived from the sepiolite mineral.

The adsorption of lead on the sepiolite samples conformed to the pseudo-second-order kinetic model. The adsorption isotherm fit the Langmuir model, with monolayer adsorption capacities of 60.6 and 90.1 mg g–1 for Sep and MagSep, respectively. The thermodynamic parameters showed that the adsorption of Pb2+ on the sepiolite samples was endothermic (ΔH° > 0) and adsorption occurred spontaneously (ΔG° < 0). Pb2+ was adsorbed successfully on the sepiolite samples, as determined from the XRF, FTIR and SEM-EDS analyses. The results revealed that MagSep can be considered to be a potentially better adsorbent than Sep for Pb2+ remediation from aqueous solutions, and with the use of an external magnetic field it can be separated easily from aqueous solutions without releasing extra toxic pollutants into the environment. It can therefore be recommended that the use of MagSep prepared from Eskişehir sepiolite be further investigated in terms of its adsorption capacity regarding other heavy metals.

The authors thank Prof Dr Yusuf Kağan Kadıoğlu for supplying the raw sepiolite samples, Earth Sciences Application and Research Center (YEBİM) at Ankara University for the XRD, XRF and ICP-OES analysis support, the Institute of Nuclear Sciences at Ankara University for SEM-EDS analyses, Prof Dr Hande Çelebi at Eskisehir Technical University for the FTIR analysis support, Middle East Technical University (METU) Central Laboratory for the ζ-potential, VSM and BET analysis support and Superconductor Technologies Application and Research Center at Ankara University for the particle-size analysis support and for providing the permanent magnet.

Available upon request.

This study was partially supported by Ankara University via project number BAP-21L0405003.

The authors declare none.

None.

Conceptualization: G.Ö. Çakal; Methodology: G.Ö. Çakal; Formal analysis and investigation: O. Uygun, A. Murat; Writing – original draft preparation: O. Uygun, A. Murat, G.Ö. Çakal; Writing – review and editing: G.Ö. Çakal; Funding acquisition: G.Ö. Çakal; Supervision: G.Ö. Çakal.

This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.