Removal of Pb(II) Ions from Aqueous Solution Using Modiﬁed Starch

: In this study, two types of modiﬁed cassava starch samples (MCS and MWS) prepared from commercially available native cassava starch (NCS) and native cassava starch extracted using the wet method (NWS) were investigated for the removal of Pb(II) ions from aqueous solutions. MCS and MWS samples were synthesized under acidic conditions using Pluronic 123 as the structure-directing agent and tetraethylorthosilicate (TEOS) as the chemical modifying agent. Modiﬁed starch samples were characterized using Fourier transform infrared (FTIR) spectroscopy, thermogravimetric analysis (TGA), X-ray Diffraction (XRD), and a nitrogen (N 2 ) gas adsorption–desorption analyser. MCS and MWS showed enhanced thermal stabilities upon acid hydrolysis and chemical modiﬁcation. The effects of contact time and initial Pb(II) concentration were studied through batch adsorption experiments. Adsorption kinetics followed the pseudo-second-order kinetic model. The equilibrium adsorption data were analysed and compared by the Langmuir and Freundlich adsorption models. The coefﬁcient correlation (R 2 ) was employed as a measure of the ﬁt. The Langmuir model ﬁtted well with equilibrium adsorption data, giving a maximum Pb(II) adsorption capacity of 370.37 and 294.12 mg/g for MWS and MCS, respectively. Modiﬁed samples exhibited a higher desorption efﬁciency of over 97%. This study demonstrated that modiﬁed starch could be utilized for Pb(II) removal from industrial wastewater.


Introduction
With the rapid industrialization and urban development across the globe, excessive release of heavy metals into the environment has significantly affected soil, plants, aquatic life, animals, and humans [1][2][3][4]. High toxicity, carcinogenicity, and non-biodegradability of Several studies have been previously investigated on the effective removal of Pb(II) ions using chemically modified starch. For instance, Awokoya and coworkers [35] reported a maximum Pb(II) adsorption capacity of 110.64 mg/g using succinylated starch. Kweon et al. [36] showed the effective removal of 48.27 mg/g of Pb(II) using oxidized starch. Guclu and coworkers [37] displayed a maximum removal of 112.96 mg/g of Pb(II) by starch graft polyacrylic acid. Xu et al. [38] studied Pb(II) ion removal by cross-linked amphoteric starch with an absorption capacity of 152.74 mg/g. Separate studies conducted by Guo et al. [39] with cross-linked phosphate carbonate starch, Soto et al. [40] with starch esters, and Koyla et al. [41] with hydroxyethyl starch-graft polyacrylamide exhibited the maximum Pb(II) removal capacity of 316. 47, 25.16, and 103.6 mg/g, respectively. Iran and coworkers [42] achieved a maximum Pb(II) adsorption capacity of 430 mg/g with polyethylene-graft poly(acrylic acid)-co-starch/organic-montmorillonite hydrogel composite. Generally, the modified starch adsorbents remove Pb(II) ions via chelation and ionic interactions.
In this study, we report the removal of Pb(II) ions from an aqueous solution using two types of modified starch samples prepared from commercially available cassava starch and native cassava starch extracted in the lab. To the best of our knowledge, this is the first attempt to modify starch with improved surface properties using Pluronic P123 and tetraethylorthosilicate (TEOS) as a structure-directing agent and a chemical modifying agent. Moreover, the synthesis process is simple, environmentally benign, low cost, and less labor intensive.
Standard Pb(II) solutions were prepared using PbNO 3 . The pH value of all Pb(NO 3 ) 2 solutions was adjusted to 4.4 to avoid the precipitation of Pb at higher pH (pH > 4.8) [16].
Two cassava starch types, (i) commercially available native cassava starch (food-grade) purchased from Vilaconic joint-stock company, Vietnam, and (ii) native cassava starch extracted using a wet method, were used for modification.

Extraction of Cassava Starch
The cassava starch extraction process was conducted according to the wet method described by Benesi et al. [43]. Initially, cassava roots were peeled, washed, and disintegrated into 1 cm cubes. After that, the cassava cubes were pulverized in a high-speed blender for 5 min. Then, the resulted pulp was suspended in ten times its volume of water and stirred for 5 min. Next, the pulp was filtered using a double-fold cotton cloth, and the filtrate was allowed to stand for 2 h for the starch granules to settle. The top liquid was then decanted and discarded. After that, water was added to the sediment, and the mixture was stirred again for 5 min. Finally, the extracted cassava starch was dried at 65 • C for 3 h. The moisture and fiber content of the cassava starch was determined using the Association of Official Analytical Chemists (AOAC) 920. 36 (2000) and 962.09 (2005) standards [44,45]. The amylose content of native starch was determined by the blue value method described by Stawski [46]. The gelatinization temperature was measured using differential scanning calorimetry (DSC, 131-Evo, Setaram Instrumentation, Caluire, France). The samples were heated from 25 to 140 • C with the scan rate at 10 • C/min. The composition of the starch samples used for the chemical modification is depicted in Table 1. All the analyses were carried out in triplicate.

Preparation of Adsorbents
Modified starch was prepared under acidic conditions using Pluronic 123 (poly(ethylene oxide) 20 -poly(propylene oxide) 70 -poly(ethylene oxide) 20 [-(EO) 20 -(PO) 70 -(EO) 20 -]) as a structure-directing agent and TEOS as the chemical modifying agent according to a slightly modified method described elsewhere by Gunathilake et al. [47]. Initially, 1 g of cassava starch purchased from Vilaconic joint-stock company, Vietnam, was dissolved in 20 mL of 0.5 M HCl under vigorous stirring at 150 rpm while maintaining the temperature at 45 • C for 3 h. Simultaneously, 2 g of Pluronic 123 was dissolved in 125 mL of 0.5 M HCl under rapid stirring at room temperature for 3 h. After that, 4.43 cm 3 of TEOS was dissolved in 50 mL of 0.5 M HCl under stirring at 150 rpm at room temperature for 2 h. Next, the acidic solution, which contains dissolved Pluronic 123, was heated up to 40 • C while stirring at 150 rpm. The TEOS solution was then added to the beaker with Pluronic 123. The mixture was left under rapid stirring at 40 • C for 3 h. Then, the starch solution was added to the TEOS/Pluronic 123 mixture. (Note that the starch solution was neutralized by adding 0.5 M NaOH before the addition). The final mixture was left under vigorous stirring for 6 h at 40 • C. Subsequently, the solution was hydrothermally treated in YAMATO DS 60 drier at 105 • C for 24 h until a powdered mixture was obtained. The powder was mixed with an HCl-ethanol solution for 12 h at 35 • C and subsequently filtered through 8 µm filter paper to remove the template. Finally, the power collected on the filter paper was dried at 80 • C for 2 h. A similar procedure was followed for modifying cassava starch extracted by the wet method. The native and modified starch samples were named during the experiments, as mentioned below.  All four samples (NCS, NWS, MCS, and MWS) were characterized by N 2 adsorption and desorption measurements. The specific surface area and porosity of the starch particles were examined using the nitrogen gas adsorption-desorption analyzer. A Brunauer-Emmet-Teller (BET) model was used to calculate the specific surface area. The monolayer adsorption was evaluated using the linear form of the BET equation. The BET surface area (S BET ), singlepoint pore volume (V sp ), the volume of fine pores (V mi ), and the total pore volume (V t ) were measured using the nitrogen adsorption analyzer. Nitrogen adsorption isotherms were measured at −196 • C on an ASAP 2010 volumetric analyzer (Micromeritics, Inc., Norcross, GA, USA). Prior to the adsorption measurements, all samples were outgassed under vacuum at 110 • C for 2 h [48].
The single-point pore volume (V sp ) was estimated from the amount N 2 gas adsorbed at a relative pressure (p/p 0 ) of~0.98. The pore size distributions (PSD) were calculated using adsorption branches of nitrogen adsorption-desorption isotherms by the improved KJS method calibrated for cylindrical pores [49]. The BET surface areas (S BET ) were calculated from the N 2 adsorption isotherms in the relative pressure range of 0.05-0.2 using a crosssectional area of 0.162 nm 2 per nitrogen molecule.

Fourier Transforms Infrared Spectroscopy (FTIR)
FTIR spectra of modified cassava starch samples were obtained using a Nicolet iS10FTIR instrument (Thermo Scientific, Waltham, MA, USA). Samples were analyzed by transmission mode after preparing KBr pellets with FTIR grade KBr (99%). Spectra were collected in the wavenumber range of 500 to 4000 cm −1 . For each sample, 4 scans were taken at a resolution of 4 cm −1 .

Thermogravimetric Analysis (TGA)
Thermal degradation of modified cassava starch samples was evaluated using a TGA 5500 (TA instruments, New Castle, DE, USA) under the nitrogen atmosphere. Platinum crucibles were of 100 µL were used with a heating rate of 10 • C/min at 25 to 650 • C. Thermal degradation temperature was calculated using the TRIOS software (4.4.0 version, TA instruments, New Castle, DE, USA).

X-Ray Diffraction (XRD)
The X-ray diffraction patterns of the starch samples were obtained using a Bruker D8 Advanced Eco Powder X-ray Diffraction system. The XRD spectra were recorded over an angular range (2θ) of 5 to 45 • with a continuous scanning at scan rate = 0.02 • /min.

Adsorption Kinetic Studies
Adsorption kinetic studies were performed using 50 mL of aqueous Pb(II) solutions with an initial concentration of 100 mg/L. The pH value of aqueous solutions was maintained at 4.4. In all experiments, 0.05 g of the adsorbent was used under similar experimental conditions with a stirring speed of 150 rpm at 22 • C. The experiments were triplicated for each adsorbent. An aliquot of 0.2 mL was taken at predetermined intervals before adding the adsorbent (t = 0) and after adding the adsorbent (t = 4-120 min). From 4 to 20 min, the samples were taken at 4 min intervals, while from 20 to 30 min, the samples were taken at 5 min intervals. From 30 to 120 min, the samples were taken at 10 min intervals. After filtration, the concentrations of heavy metal ions in the aqueous solutions were determined using atomic absorption spectroscopy (AAS). The adsorption capacity was calculated from the following Equation (1): where q t (mg/g) is the adsorption capacity of the adsorbent, C 0 (mg/L) is the initial concentration of Pb(II) solution, C t (mg/L) is the concentration of Pb(II) at a given time, V (L) is the volume of Pb(II) solution, and m (g) is the mass of the adsorbent.

Equilibrium Adsorption Kinetics
All adsorption experiments were conducted batch-wise. A series of Pb(II) solutions with different initial concentrations (50,100,150,200,250, 300 mg/L) were prepared to determine the equilibrium adsorption behaviour. The equilibrium kinetics were followed using the procedure mentioned before in a 50 mL of Pb(II) solution with 0.05 g of adsorbent under vigorous stirring at 150 rpm at 22 • C for 12 h. After the filtration, the concentrations of Pb(II) ions in the aqueous solutions were determined by AAS. All the experiments were triplicated.
The adsorbed heavy metal ions per unit mass of the solid adsorbent at equilibrium (q e ) were determined according to the following Equation (2): where q e qe(mg/g) is the equilibrium adsorption capacity of the adsorbent, C 0 (mg/L) is the initial concentration of Pb(II) solution, C e (mg/L) is the equilibrium concentration of Pb(II), V (L) is the volume of Pb(II) solution, and m (g) is the mass of the adsorbent.

Desorption Experiments
Desorption studies were carried out in a batch system using Pb(II) ion loaded adsorbents immediately after the adsorption experiments. The regeneration experiments were carried out only for the two modified starch samples. A 0.05 g of adsorbent was mixed with 100 mL of 0.1 M HCl solution for 30 min at room temperature (RT). The desorption was followed in each cycle at every 10 min intervals. The concentration of Pb(II) after desorption was measured using AAS, as previously reported. The desorption efficiency (DE) was calculated according to the following Equation (3) [9]: where DE (%) is the desorption efficiency, C t (mg/L) is the concentration of lead ions in the desorption solution at time t (min), V is the volume of the desorption solution, and m 0 (g) is the amount of Pb(II) adsorbed.

Synthesis of Modified Starch
Synthesis of the modified starch involves the hydrolysis and condensation steps, as shown in Schemes 1 and 2. The structure-directing agent, Pluronic 123, is a block-copolymer containing hydrophilic ethylene oxide (EO) and hydrophobic propylene oxide (PO) parts. During the synthesis, both starch and TEOS are hydrolyzed under acidic conditions; see Scheme 1. Due to the hydrophilic and hydrophobic nature of P123, it forms micelles. Hydrophilic part of the P123 and condensation product formed from the reaction between the hydrolyzed TEOS and starch have weak electrostatic interaction through hydrogen bonding. The micelles formed by P123 are subsequently removed with acidic ethanol to enhance porosity, as illustrated in Scheme 2.

N 2 Adsorption Studies
The N 2 adsorption-desorption isotherms of two native (NCS and NWS) and two mod- Our results suggest that the surface properties of the modified starch significantly enhanced after the chemical modification (see Table 2). For instance, the modified starch extracted by the wet method (MWS) sample shows the highest specific surface area and total pore volume of 5.8 (±0.1) m 2 /g and 0.022 (±0.001) cm 3 /g. Native commercially available starch (NCS) exhibits the least specific surface area and total pore volume of 1.7 (±0.1) m 2 /g and 0.007 (±0.001) cm 3 /g (see Table 2). The structural parameters of the native and modified starch samples calculated based on the N 2 adsorption-desorption data are depicted in Table 2.
V sp -single point pore volume calculated at the relative pressure of 0.98; V mi -volume of fine pores (micropores below 2 nm) calculated by integration of the PSD curve up to 2 nm; S BET -specific surface area calculated from adsorption data in relative pressure range 0.05-0.20; V tot -total pore volume calculated by integrating PSD in the entire relative pressure range.

Fourier Transforms Infrared Spectroscopy (FTIR)
FTIR studies were conducted on all starch samples. Figure 2 shows the FTIR spectra of both native (NCS and NWS) and modified starch (MCS and MWS). All samples show broad O-H stretching bands at 3437-3397 cm −1 [50], and a band around 2929 cm −1 corresponds to the stretching vibration of the C-H bond from the glucose units; see Figure 2. Both MCS and MWS samples display a peak at 1728 cm −1 that is ascribed to the C=O group, suggesting the oxidation of alcohol to aldehyde during the hydrolysis process. The absorbance band at 1647 cm −1 that appeared in all samples is attributed to the bending vibration of O-H of absorbed water. The peaks at 1469 and 1381 cm −1 are for CH 2 symmetric scissoring and C -H symmetric bending vibrations. Two bands at 1172 and 1104 cm −1 correspond to the C-O-C asymmetric stretching and C-O stretching vibrations, respectively [51]. The peak at 957 cm −1 is assigned to the skeletal mode vibration of α-(1-4) glycosidic linkage. Both MCS and MWS samples exhibit a sharp 957 cm −1 peak due to the hydrolysis of glycosidic linkages. Moreover, both MCS and MWS samples do not show the IR bands at 1172 and 1014 cm −1 due to the breakdown of α,1-4 and α, 1-6 glycosidic bonds upon acid hydrolysis [52].

Thermogravimetric Analysis (TGA)
Thermal stability of all four starch samples (NCS, NWS, MCS, and MWS) were investigated in flowing nitrogen using high-resolution thermogravimetry. Figure 3 exhibits the thermogravimetric (TG) and corresponding differential thermogravimetric (DTG) profiles of all starch samples. The TG profiles of NWS and NCS samples (see Figure 3a,b) show two distinct weight loss regions in the temperature ranges of 25-100 • C and 250-580 • C. The former event represents a weight loss of~12% (w/w) corresponding to the evaporation of physically adsorbed water. The latter event is attributed to the thermal decomposition of starch with a weight loss of ∼63% (w/w). The DTG profiles of the NWS and NCS display the thermal degradation temperature of starch at 330 • C. The TG and DTG profiles of MWS and MCS samples are shown in Figure 3c,d. Both MWS and MCS samples exhibit three weight loss regions at 25-130, 130-220, and 250-580 • C, respectively. The first thermal event refers to the evaporation of physically absorbed water. The second and third events are ascribed to the thermal decomposition of the residual triblock copolymer template and starch, respectively. The residual triblock copolymer template decomposed at 167 • C for both MWS and MCS samples. However, the thermal decomposition of the starch in the modified samples occurred at 357 and 378 • C for MWS and MCS, respectively. Both MWS and MCS samples show an increase in their thermal stability after the hydrolysis and chemical modification. One of the two major reasons for the enhanced thermal stabilities of MWS and MCS is the increased crystallinity, as the acid hydrolysis mainly occurs in the amorphous regions of the starch. The other reason is the incorporation of silica into the polymer structure, which provides more rigidity.      In order to study the adsorption kinetics of Pb(II) ions, the kinetic parameters were investigated using Lagergren's first-order kinetics model and Ho's pseudo-second-order kinetic model.

Pseudo-First Order Kinetic Model
The pseudo-first-order kinetics of the adsorption of Pb(II) onto starch was followed using the Lagergren first-order kinetic model; see Equation (4) [34].
The first-order adsorption kinetic model is given below: Lagergren's pseudo-first-order model can be expressed as in Equation (5) by plotting the linear form of Equation (5), log(q e − q t ) = log(q e ) − k 1 t 2.303 (5) where k 1 (min −1 ) is the equilibrium rate constant of pseudo-first-order adsorption, q e (mg·g −1 ) is the amount of Pb(II) ions adsorbed at equilibrium, q t (mg·g −1 ) is the amount of Pb(II) ions adsorbed at time t (min). The data were fitted to the Lagergren pseudo-first-order model; see Equation (5). Figure S1 (Supporting Information) shows the plots of the linearized form of pseudo-firstorder model (log(q e − q t ) vs. t) for the adsorption of Pb(II) ions onto NWS, NCS, MWS, and MCS samples. The first-order rate constants of MCS and MWS samples are higher than their respective native starch samples, NCS and NWS.
The applicability of Lagergen's pseudo-first-order model was determined by calculating the correlation coefficient, R 2 . Table 3 summarizes the pseudo-first-order rate constants and the correlation coefficients (R 2 ) of first-order kinetic plots. Generally, the linearity of the plots indicates the applicability of the kinetic model of interest. As shown in Table 3, R 2 was found to be in the range of 0.92-0.98 for all first-order kinetic plots. Our results suggest that the adsorption process is not in good agreement with the pseudo-first-order kinetic model. Table 3. First-order rate constants (k 1 ) and the correlation coefficients (R 2 ) of first-order kinetic plots.

Sample
First-Order Rate Constant (k 1 ), min − Ho's pseudo-second-order model was employed to study the second-order adsorption kinetics.
The pseudo-second-order kinetic model is given by Equation (6): In the linear form, as shown in Equation (7), where k 2 (g·mg −1 min −1 ) is the rate constant for the pseudo-second-order adsorption model, q e (mg·g −1 ) is the amount of Pb(II) ions adsorbed at equilibrium, q t (mg·g −1 ) is the amount of Pb(II) adsorbed at time t (min). The plot between t/q t and t was recorded to determine the rate constant for the pseudo-second-order adsorption model. The concentration adsorption rate h (mg·g −1 ·min −1 ) can be calculated according to Equation (8) [13].
The plots of the linearized form of pseudo-second-order model (t/q t vs. t) are depicted in Figure S2a,b, Supporting Information. Table 4 summarizes the kinetic parameters calculated from the pseudo-second-order model and the correlation coefficients (R 2 ) for all plots. Data fitted well to the pseudo-second-order model with the correlation coefficients over 0.98 (R 2 ≥ 0.98); see Table 4. The linearity of these plots indicates the applicability of this model. Therefore, our data suggest that the adsorption process occurs via a chemisorption mechanism. As can be seen from Table 4, the MWS sample shows the maximum concentration adsorption rate (h) of 5.52 mg·g −1 ·min −1 .

Effect of Initial Concentration on Adsorption Capacity
The effect of initial Pb(II) concentration on the adsorption capacity was conducted with varying initial concentrations of Pb(II) (50,100,150,200,250, 300 mg/L) and a fixed adsorbent dose of 1 g/L. Figure 6a,b show the plots of adsorption capacity against the initial Pb(II) concentration for NCS, NWS, MCS, and MWS samples. As can be seen from Figure  5a,b, the amount adsorbed increased with increasing Pb(II) ion concentration, indicating the availability of actives sites on the adsorbate. This also suggests that the adsorption process is highly concentration dependent. The amount of Pb(II) ions adsorbed by NCS and NWS samples is significantly lower as compared to their respective MCS and MWS samples. MCS and MWS samples show a maximum absorption capacity of 216.54 and 218.74 mg/g, respectively. Our results reveal that the starch modification immensely contributed to the higher adsorption capacities.

Adsorption Isotherms
The equilibrium adsorption isotherms are commonly considered to understand the mechanism of the adsorption. The most widely used Langmuir and Freundlich isotherms were employed to study the equilibrium of the adsorption system. The adsorption isotherms were constructed at a fixed adsorbent dosage (1 g/L) by varying initial concentrations of Pb(II) ions (50-300 mg/L).

Langmuir Model
Langmuir isotherm models the monolayer coverage of the adsorption surfaces [34]. The Langmuir model is applicable for homogeneous adsorption studies based on the following assumptions: (i) all the adsorption sites are assumed to be identical, and all sites are energetically and sterically independent of the adsorbed quantity; (ii) each site retains only one molecule of the given compound; (iii) adsorbent and intermolecular forces decrease rapidly with the distance from the adsorption surface [29].
Langmuir equation is given below: In the linear form given by Equation (10), where K L (L mg −1 ) is the Langmuir adsorption constant related to the energy of adsorption, which reflects the affinity between the adsorbent and adsorbate, q m (mg·g −1 ) is the maximum adsorbed quantities, C e (mg·L −1 ) is the equilibrium concentration, and q e (mg·g −1 ) is the amount of lead adsorbed at equilibrium. An essential parameter related to the Langmuir model can be expressed in terms of dimensionless constant separation factor of equilibrium parameter R L , is represented by Equation (11), which is used to determine if a certain sorption system is "favourable" or "unfavourable". The determining factors are as follows: (i) unfavourable (R L > 1); (ii) Linear (R L = 1); (iii) favourable (R L < 1) and (iv) unfavourable (R L = 0) [31].
where C i (mg·L −1 ) is the initial concentration of the lead ion solutions. The Langmuir isotherm exhibited a better fit to the experimental data with higher correlation coefficients for all starch samples (R 2 ≥ 0.98), as given in Table 5 and Figure S3 (Supporting Information). Therefore, the adsorption mechanism is well described by the monolayer Langmuir model on a homogeneous surface. The adsorption capacities of the MCS and MWS were remarkably higher as compared to their unmodified counterparts; see Table 5. The maximum adsorption capacities of MCS and MWS are found to 294.12 and 330.37 mg/g, respectively. Table 5 summarizes the Langmuir isotherm constants for the adsorption of Pb(II) ions. The R L value calculated from the Langmuir model is in the range 0-1 for all samples; see Table S1 (Supporting Information). Moreover, the R L value decreases with increasing initial Pb(II) concentration, which indicates the favourable adsorption of Pb(II) onto starch samples; see Table S1 (Supporting Information).

Freundlich Model
The Freundlich model is among the most widely used models to analyse the equilibrium data related to adsorbents. Freundlich equation is employed to describe the non-ideal multilayer adsorption on to the heterogeneous surfaces [44].
The Freundlich isotherm is given below: The linear form of the Freundlich isotherm can be expressed as log(q e ) = log(K F ) + N log(C e ) (13) where is the Freundlich constant being indicative of the strength of the adsorption bond, N is the heterogeneity factor as an indicator of adsorption effectiveness, C e (mg·L −1 ) is the equilibrium concentration, and q e (mg·g −1 ) is the amount of Pb(II) ions adsorbed at equilibrium. Furthermore, K F predicts the quantity of Pb(II) ions adsorbed per gram of composite at the equilibrium concentration. If N < 1, the bond energies increase with the surface density while N > 1, bond energies decrease with the surface density. Moreover, if N = 1, all surface sites are equivalent. If the value of N is in the range of 0 and 1, it indicates favourable sorption [29].
The equilibrium data were fitted to the Freundlich model; see Figure S4 (Supporting Information). Table 6 summarizes the Freundlich isotherm constants for Pb(II) adsorption. According to the data shown in Table 6, both MCS and MWS samples show larger K F values over NCS and NWS samples, indicating their higher adsorption capacity and affinity for Pb(II) ions. All calculated N values are within the range of 0 to 1, suggesting a favourable sorption process; see Table 5. However, the correlation coefficient (R 2 ) of the Freundlich isotherm was found to be less than or equal to 0.98 ((R 2 ≥ 0.98); see Table 6. Comparing the correlation coefficient (R 2 ), the equilibrium data for the adsorption of Pb(II) ions onto the starch samples followed the Langmuir model more than the Freundlich model. Therefore, homogeneous monolayer adsorption is more favourable over heterogeneous multilayer adsorption.

Lead Adsorption Mechanism
The absorption behaviour of Pb 2+ ions onto modified starch can be explained using Pearson's hard-soft acid-base (HSAB) principle. The pendant hydroxyl groups act as the hard Lewis bases, which show a greater affinity towards moderately hard Lewis acid, Pb 2+ . The lone pairs present in the oxygen atom donates electrons to the Pb 2+ ions resulting in complexion between the Pb 2+ ions and hydroxyl groups on the modified starch, as shown in Scheme 3. Note that it is also possible to form similar type of structure between uncondensed Si-OH functional groups and Pb 2+ as shown in Scheme 3.

Desorption Studies
An imperative characteristic of an adsorbent is the desorption efficiency. Table 7 presents the desorption efficiencies of the modified starch samples (MCS and MWS). Both MCS and MWS exhibit high desorption efficiencies over 97% within 30 min. Our data suggest that the adsorption process is reversible, and hence the adsorbents can be recycled. Moreover, the recovered Pb(II) ions during the desorption process could be used for other applications.

Comparison with Other Studies
Several studies have been previously reported on the removal of Pb(II) ions from aqueous solutions using various adsorbents, including starch-based materials. Table 8 presents a comparison of the maximum Pb (II) ion adsorption capacities for different adsorbents reported in the literature. Interestingly, the modified starch samples studied in this study show comparatively larger adsorption capacities; see Table 8. Polyethylene-graftpoly (acrylic acid)-co-starch/organomontmorillonite hydrogel composite 430 [42]

Conclusions
In this study, two types of modified starch adsorbents (MCS and MWS) were successfully prepared using Pluronic 123 as a structure-directing agent and TEOS as the chemical modifying agent. The adsorption process followed the pseudo-second-order kinetic model indicating the chemisorption mechanism. The equilibrium adsorption data fitted well to the Langmuir isotherm model with a correlation coefficient (R 2 ) over 98%. The adsorption capacities calculated from the Langmuir isotherm were 330.37 and 294.12 mg/g for MCS and MWS. The MCS and MWS samples showed the highest desorption efficiencies of over 97%. Due to the easy synthesis route, cost-effectiveness, environmental benignity, high adsorption capacities, and desorption efficiencies, modified starch materials show potential in large-scale removal of Pb(II) ions from aqueous solutions.

Data Availability Statement:
The data presented in this study are openly available in Electronic Supporting Information (ESI). Data not presented in this study are available on request from the corresponding author.