Facile Method by Bentonite Treated with Heat and Acid to Enhance Pesticide Adsorption

: In this work, simple conditions were applied to modify bentonite for the removal of pesticides from aqueous solution. Bentonite was modiﬁed in a single step as BA 0.5 (with HCl 0.5 M) and BC 500 (calcined at 500 ◦ C) and combined steps with different sequences (BA 0.5 C 500 and BC 500 A 0.5 ). These adsorbents were characterised by XRD, XRF, FT-IR, 27 Al MAS NMR, BET, NH 3 -TPD, TGA, HPLC, particle size analysis and zeta potential. Single-component adsorption with atrazine, diuron, 2,4-D and paraquat was used in aqueous solution at various pesticide concentrations, contact times and pH levels. It was found that the sequence of the treatment signiﬁcantly affected atrazine adsorption. BC 500 A 0.5 exhibited the highest efﬁciency for atrazine adsorption in a broad pH range of 3.0–9.0. Its adsorption at pH 6.0 was about 12 times greater than that of other adsorbents with an initial atrazine concentration of 50 mg L − 1 , which indicates BC 500 A 0.5 speciﬁcally for the adsorption of atrazine. In addition, for the simultaneous adsorption of all four pesticides, BC 500 A 0.5 was found to remove the maximum total amount of the pesticides, indicating that it could be used as a good multifunctional adsorbent. All modiﬁed bentonites showed similar diuron adsorption better than that of unmodiﬁed bentonite. The greatest adsorption of 2,4-D prefers BA 0.5 C 500 , occurring at pH 2–4. In the case of paraquat adsorption, all adsorbents are good at absorbing paraquat, but bentonite had the highest rate of paraquat removal, whereas BA 0.5 C 500 was found to have the lowest, and the adsorption increased with increasing pH. Furthermore, the adsorption process on the adsorbents ﬁts well with the Langmuir isotherm and pseudo-second-order kinetics models, as the thermodynamic parameters showed a spontaneous and endothermic process.


Introduction
Toxic pesticides have a significant impact on human health as well as aquatic life. Currently, it is a topic of significant concern because of the increasing use of pesticides in agriculture and domestic activities [1][2][3][4][5][6][7]. Therefore, it is necessary to find ways to limit the agrochemical substances discharged into the environment as much as possible, while also acting to eventually eliminate them completely.
Adsorption is a simple and convenient method for the removal of compounds compared to other physical, chemical, or biological technologies. It also involves low investment in terms of both initial cost and land required. Various solid substrates offer good adsorptive properties, including zeolites, activated carbon, fly ash, mesoporous silica and clay minerals, which were applied in the removal of organic contaminants from water [2,[8][9][10][11][12][13][14].
Bentonite clay minerals are available not only in Thailand, but their deposits are widespread all over the word. They are an important, low-cost source of montmorillonite and exhibit excellent adsorption capacities as well as high cation exchange capability (CEC) [15,16]. The major component of bentonite is montmorillonite, a dioctahedral smectite. Montmorillonite is a porous clay mineral consisting of a 2:1-layered structure, with alternating layers of exchangeable cations. The layers consist of octahedral alumina sheets

Characterisation
The chemical compositions of the major elements in bentonite and modified bentonite were determined by energy dispersive XRF (EDS, Oxford Instrument ED 2000, Abingdon, UK) using an Rh X-ray tube with a vacuum medium. The CEC was determined by a 9080 CEC standard method (US EPA, OSWER), in which all adsorbed cations in a bentonite sample were replaced by a common ion, such as NH 4 + , and then the amount of adsorbed common ion was determined. The interlayer spacing of the samples was determined by X-Ray diffraction patterns obtained from back-pressed powder samples recorded on a Bruker, D5005, Germany, Cu Kα radiation, scanning from 3 • to 50 • at a rate of 0.05 • /s at 35 mV and 35 mA. The framework and OH groups were confirmed by FT-IR spectroscopy (Spectrum GX, Perkin-Elmer, Waltham, MA, USA) as KBr pellets, over the range of 4000-400 cm −1 and near-IR diffuse reflection (DRIFT) in the range of 8000-4000 cm −1 . The DRIFT technique was especially appropriate in this region because no sample dilution was required. Specific surface and pore size distributions were evaluated by nitrogen gas adsorption at −196 • C using a BEL JAPAN, BELSORP-mini II, Japan instrument, with the Brunauer-Emmertt-Teller (BET) method. All samples were characterised by solid-state 27 Al CP/MAS NMR (JEOL JNM-ECA600), with a static 11 T magnetic field at a spin rate of 11 kHz. The 27 Al CP/MAS NMR spectra were recorded at a frequency of 130 MHz. The NMR reference used was Al(NO 3 ) 3 .9H 2 O with a chemical shift of 0.00 ppm. The particle size distributions (PDS) of all samples were analysed by a Horiba laser-scattering particle-size-distribution analyser (LA-950), Japan. The active site was determined using NH 3 temperature-program desorption (TPD) apparatus (BEL Japan, model BELCAT-B, Osaka, Japan). Samples were pre-treated at 300 • C for 2 h and then exposed to a flow of NH 3 /He at 100 • C for 30 min. Weakly adsorbed NH 3 was removed by flushing with He gas at 100 • C for 30 min. The NH 3 TPD profile was obtained over the range of 100 • C-600 • C at a heating rate of 10 • C/min. Desorbed NH 3 was detected using a thermal conductivity detector. TGA curves were obtained by heating the sample in an alumina70 cell from 30 to 800 • C at a heating rate of 10 • C min −1 under air atmosphere using a thermogravimetric analyser (TGA-DSC, METTLER, Germany). Zeta potential measurements were performed using a Malvern, UK. Bentonite and modified bentonite slurries were homogenised using a Branson Sonifier, UK. Sonifier employs ultrasonic vibration to disperse the dry clay particle aggregates in suspensions. Before the sonication, the amount of clay powder and deionised (DI) water was kept constant at 10 wt.%. Dilution suspensions were prepared for the zeta potential measurement with a colloidal Dynamics Zeta Probe. The concentration employed was only 1 wt.% solids. The pH was increased to approximately 12 with 1 M NaOH before commencing the yield stress measurements. Concentrated HCl (1 M) was added dropwise to decrease the pH in a stepwise manner.
Atrazine, diuron, 2,4-D, and paraquat concentration were determined using UVvisible spectrophotometry at 223, 247, 283 and 257 nm, respectively (T80 + UV-Vis spectrophotometer PG instruments, Alma Park, Leicestershire, UK). The mixture solution of atrazine, diuron, 2,4-D and paraquat concentration were determined using HPLC (Hewlett Packard series 1100, Palo Alto, CA, USA). All experimental measurements were made in duplicate. The quantity of adsorbed atrazine, diuron, 2,4-D and paraquat by bentonite and modified bentonite was calculated from the difference between the initial and final pesticide concentrations using the mass balance Equation (1): where q e is the adsorbed pesticide remaining on adsorbents (mg g −1 ) and C i and C e are the initial and final pesticide concentrations (mg L −1 ), respectively. M is the mass of adsorbent (g), and V is the volume of the pesticide solution (L).

Batch Adsorption
All adsorbents were dried overnight at 110 • C prior to commencing the adsorption study. Adsorption experiments were performed using a batch adsorption procedure with 20 mg of bentonite or modified bentonite added to 10 mL of pesticide solution, which was then kept in a shaker bath at 30 • C for 24 h (equilibrium time~12 h). The stock solution was prepared in methanol solutions; final methanol concentrations were less than 0.5% (v/v) to prevent co-solvent effects. Suspensions were centrifuged at 3500 rpm and then filtered. Each pesticide concentration in filtrates was determined by UV-Vis spectroscopy.

Effect of pH
The effect of pH on the adsorption of atrazine, diuron, 2,4-D, and paraquat by bentonite and modified bentonite was investigated over the pH range of 2.0-12.0. Pesticide solutions were prepared by dissolving the samples in distilled water. The adsorbents were exposed to the pesticide solution at a solid/liquid ratio of 20 mg/10 mL. The solution pH was adjusted by adding 0.1 or 0.01 M of HCl or NaOH solution, as appropriate. The mixture was placed in a shaker at 30 • C until it reached equilibrium, then centrifuged to separate the filtrate from the solid phase. The filtrate was analysed for residual pesticide concentration. Duplicate tests were performed in all cases.

Simultaneous Adsorption of Various Pesticides in Solutions
The aqueous adsorbate solutions used in the adsorption study were: (i) a singlecomponent (each pesticide); and (ii) a quaternary-component solution of atrazine, diuron, 2,4-D and paraquat. The initial concentration of each pesticide was 50 mg L −1 . The initial pH value was unadjusted (6.5 ± 0.5). The experiments were realised at room temperature using equal volumes of the pesticide solutions, V = 10.0 mL. The mass of adsorbent in all adsorption studies was 20.0 mg. In the quaternary-component solution of the pesticides, the adsorption capacity of individual pesticides on bentonite and modified bentonite was investigated. The quaternary-component solution with 10 mL of each pesticide concentration 50 mg L −1 was added into 20 mg of adsorbent. The adsorption experiments were performed in a shaker and then kept in a shaker for 24 h. Suspensions were centrifuged at 3500 rpm and then filtered. Each pesticide concentration in filtrates was determined by UV-Vis spectroscopy. The quaternary-component solution containing all the investigated pesticides (atrazine, diuron, 2,4-D and paraquat) was determined by HPLC (Hewlett Packard series 1100) with a diode array detector and UV detector. The ZORBAX Eclipse ODS C18 column (250 mm × 4.6 mm, particle size 5 µm) was equipped with the HPLC system and the UV detector wavelength was set at 201, 223, 254 and 283 nm. Acetonitrile (A) and water pesticide (2% phosphoric acid) (B) were used as mobile phases. The gradient program of HPLC was as follows: 0−13 min 100% B; 3−15 min 60% A and 40% B; 15−18 min 100% (A) and 20−30 min 100% A. The flow rate was set at 0.9 mL min −1 .

Adsorption Isotherm
Adsorption isotherms for atrazine, diuron, 2,4-D and paraquat were acquired by batch equilibration using 20 mg of bentonite or modified bentonite in 10 mL of aqueous pesticide solution at various concentrations (10-120 mg L −1 ) at 6.0 ± 0.5. The pH values were checked before and after adsorption measurement to ensure a small change of pH values. After equilibration (12 h), solid and liquid phases were separated by centrifugation. Atrazine, diuron, 2,4-D, and paraquat concentration were determined spectrophotometrically by measuring absorbance at 223, 247, 283, and 257 nm, respectively. Data obtained from the adsorption tests were used to calculate the adsorption capacity, q e (mg g −1 ). The adsorption isotherms were evaluated by applying the Langmuir and Freundlich models.

Kinetic and Thermodynamics Adsorption Study
Kinetic adsorption experiments were performed at 30 • C by contacting the adsorbents with pesticide solutions at a solid/liquid ratio of 20 mg/10 mL. In the first phase of the study, the mixtures were shaken and the supernatants were collected at the following times: 2, 4, 6, 8, 10, 30, 60, 90, 120, 180, 240, 360, 720, 1080 and 1440 min. Afterward, supernatants were centrifuged and measured spectrophotometrically at 223, 247, 283 and 257 nm. All of the batch tests were performed in duplicate. Average values were used in the calculation. Deviations were less than 2%. In order to determine the adsorption properties at different temperatures, the above experiments were repeated at 303, 313 and 323 K. To evaluate the adsorption thermodynamics of pesticides onto the adsorbent, enthalpy (∆H), entropy (∆S) and free energy (∆G) changes explained the thermodynamic data for adsorption process, which were calculated by Equations (4) and (5).  Table 1 shows the effects of acid and heat treatment on the chemical and CEC properties of the adsorbents. Acid treatment has a significant effect on chemical composition. The weight percentage of oxide metal of Na + , Mg 2+ , K + , Ca 2+ , Fe 3+ , and of Al 3+ decreases significantly during treatment with acid in both single-and two-step modified samples, such as BA 0.5 , BC 500 A 0.5 , and BA 0.5 C 500 resulting from the effects of acid leaching. When comparing the SiO 2 /Al 2 O 3 ratios of BC 500 A 0.5 and BA 0.5 C 500 , it shows that BA 0.5 C 500 , treated with acid prior to heat treatment has a higher SiO 2 /Al 2 O 3 ratio (see Table 1), probably due to the calcined sample (BC 500 ) with larger mean particle sizes than unmodified bentonite before further treatment with acid (as mentioned in Topic 3.1.8). There are various methods available to determine CEC. Most approaches involve the complete exchange of naturally occurring cations in a clay mineral with a cationic species such as ammonium ion, followed by the removal of excess substitution ion and the determination of the amount retained in the clay interlayer [15]. It was found that CEC values for acid and heat-treated samples are lower than those for bentonite (see Table 1). The decrease in CEC value with acid treatment results from cationic leaching, while low CEC values resulting from heat treatment may be explained by a charge reduction as the result of the migration of small radii cations into vacant octahedral sites and the formation of covalent bonds after heating [31]. Another explanation is that the cation becomes more strongly bound to the surface during calcination, so the exchange by NH 4 + becomes more difficult. However, the extent of charge reduction depends on the nature of the cation and the heating temperature.

XRD Analysis
The swelling of bentonite is a consequence of its hydrated structure, in which cations are located in the interlayer space and solvated by water molecules adsorbed from the atmosphere [31]. XRD patterns of bentonite (34.8% montmorillonite) ( Figure 1) show characteristic diffraction peaks at 20 = 6.3, 19.8 and 34.9. Quartz (Q) and cristobalite (C) impurities are observed in the patterns. The d 001 reflection at 2θ = 6.30 corresponds to d-spacing = 14.02 Å, which is an interlayer comprising the mono-and bi-layers of water molecules [32,33]. A decrease in d-spacing from calcination was caused by the removal of interlayer water and the migration of exchangeable cation [31]. Moreover, d-spacing decreases from the acid treatment result from the removal of some cations. BC 500 shows d-spacing = 9.65 Å, while BA 0.5 shows d 001 = 13.00 Å, as shown in Table 2, indicating that heat treatment significantly affects basal spacing. All adsorbents, except BC 500 , show small changes in d-spacing after the pesticide adsorptions, excluding paraquat, suggesting that the pesticides probably bind to the active sites of the face and edge, whereas the pesticides adsorptions may also involve the interlayer space for BC 500 . In the case of d-spacing for cationic pesticide (paraquat), it shows a similar value (12.05-12.38 Å), except BA 0.5 C 500 , suggesting that hydrated paraquat possibly exchanges in the interlayer space with a smaller size compared to interlayer water.
water and the migration of exchangeable cation [31]. Moreover, d-spacing decreases from the acid treatment result from the removal of some cations. BC500 shows d-spacing = 9.65 Å, while BA0.5 shows d001 = 13.00 Å, as shown in Table 2, indicating that heat treatment significantly affects basal spacing. All adsorbents, except BC500, show small changes in dspacing after the pesticide adsorptions, excluding paraquat, suggesting that the pesticides probably bind to the active sites of the face and edge, whereas the pesticides adsorptions may also involve the interlayer space for BC500. In the case of d-spacing for cationic pesticide (paraquat), it shows a similar value (12.05-12.38 Å), except BA0.5C500, suggesting that hydrated paraquat possibly exchanges in the interlayer space with a smaller size compared to interlayer water.

Thermogravimetric Analysis
The weight loss observed between 50 • C and 150 • C and between 150 • C and 400 • C is due to the evaporation of water adsorbed on the adsorbents and probably due to hydrated water bound tightly to cations, respectively. The amounts of water contained are shown in Table 3 and the supplementary information in Figure S1. This result shows that BC 500 and BA 0.5 C 500 possess low hydrophilic properties due to low water content. Above 500 • C, the weight loss is traceable to the dehydration of some trapped water molecules in the interstices of the clay layers and due to dehydroxylation of the mineral, which is accompanied by a change in the crystal structure [34].

FT-IR Analysis
The middle-IR and near-IR spectra of adsorbents are shown in Figures 2 and 3, respectively. A characteristic bentonite stretching band at 3626 cm −1 is attributed to OHdioctahedral smectite. These spectra confirm that the modified bentonite samples are still characteristic of bentonite. Bands at 3420 cm −1 and 1636 cm −1 correspond to the stretching and bending vibrations of hydrogen-bonded water molecules [35]. After calcination, the spectra show a decrease in the intensities of the Al-Al-OH band at 913-920 cm −1 and the Al-Mg-OH band at 840 cm −1 for BC 500 and BA 0.5 C 500 samples, respectively, suggesting the occurrence of dehydroxylation during calcination and the partial leaching of octahedral Al cations in acid-treated samples [19,28]. Changes in the tetrahedral sheets are reflected in the position and shape of the Si-O stretching band [35]. There is a slight decrease in the tetrahedral Si-O band intensity at 1039 cm −1 for modified samples. We assigned this to Si-O vibrations within a three-dimensional amorphous phase formed during the acid and thermal treatments [35]. Two bands at 520 and 469 cm −1 correspond to bending vibrations of Al-O-Si and Si-O-Si, respectively. The band at 520 cm −1 is the most sensitive band in the presence of any remaining Al 3+ species in the octahedral sheets and two samples of BA 0.5 and BA 0.5 C 500 obviously show those with a broader band. The band at 622 cm −1 results from coupled Al-O and Si-O out-of-plane vibrations [36].
NIR spectra were recorded to determine the Si-OH groups in the adsorbent ( Figure 3). The bands at 4533 cm −1 and 7078 cm −1 are assigned to the combination mode and first overtones (2ν OH ) of structural OH groups, respectively, while the band at 5240 cm −1 corresponds to the combination mode of water molecules [26,28]. BA 0.5 , BA 0.5 C 500 , and BC 500 A 0.5 shows a new band at 7319 cm −1 assigned to the first overtone for silanol (2ν Si-OH ), confirming the formation of Si-OH groups during treatment with acid [35]. Appl. Sci. 2021, 11, x FOR PEER REVIEW 8 of 23 NIR spectra were recorded to determine the Si-OH groups in the adsorbent ( Figure  3). The bands at 4533 cm −1 and 7078 cm −1 are assigned to the combination mode and first overtones (2νOH) of structural OH groups, respectively, while the band at 5240 cm −1 corresponds to the combination mode of water molecules [26,28]. BA0.5, BA0.5C500, and BC500A0.5 shows a new band at 7319 cm −1 assigned to the first overtone for silanol (2νSi-OH), confirming the formation of Si-OH groups during treatment with acid [35].   NIR spectra were recorded to determine the Si-OH groups in the adsorbent ( Figure  3). The bands at 4533 cm −1 and 7078 cm −1 are assigned to the combination mode and first overtones (2νOH) of structural OH groups, respectively, while the band at 5240 cm −1 corresponds to the combination mode of water molecules [26,28]. BA0.5, BA0.5C500, and BC500A0.5 shows a new band at 7319 cm −1 assigned to the first overtone for silanol (2νSi-OH), confirming the formation of Si-OH groups during treatment with acid [35].  The NH 3 -TPD profile provides information on the strength and density of adsorbent surface acid sites. The NH 3 -TPD profile of pristine bentonite shows three discrete desorption maximal peaks in the low-temperature region (50-200 • C), medium temperature region (200-420 • C), and high-temperature region (420-800 • C), which are attributed to weak, medium, and strongly acidic sites, respectively [19]. Figure 4 shows NH 3 -TPD plots for bentonite and modified bentonite. All adsorbents exhibit a peak in the high-temperature region which corresponds to a strong acid site. We attribute this to the desorption of NH 3 from Al on the surface of the adsorbent. The greatest density (3.071 mmol g −1 ) for these acid sites is seen for BC 500 A 0.5 . The NH 3 -TPD profiles for acid-activated adsorbents show additional peaks in the mid-temperature region (400-475 • C) that can be ascribed to medium acidic Si-OH 2 + sites [19].
Appl. Sci. 2021, 11, x FOR PEER REVIEW 10 of 23   3.1.6. NMR Analysis Figure 5 shows 27 Al MAS NMR spectra for all adsorbents. A major peak at 3.73 ppm in the spectrum for bentonite arises from six-coordinate Al (Al(VI)), indicating that much of the aluminium is located in octahedral environments. It is notable that the octahedral peak is split into the acid treatment performed in the final step. The two minor peaks at 68.25 Al(IV)a and 55.76 Al(IV)b are attributed to the four-coordinate tetrahedral Al Q 3 (3Si) and Q 4 (4Si) core units, respectively [37,38]. This minor peak area decreases after modification with acid. The result suggests that four-coordinate tetrahedral Al is easily removed. In the case of calcination (BC 500 ), the observed decrease in Al(IV) peak may result from some Al in the bentonite tetrahedral sheets probably being substituted by other cations such as Mg 2+ and Ca 2+ due to cation migration, which corresponds to previous reports [28,39]. In summary, these arguments account for the disappearance of two minor peaks of tetrahedral Al. A small increase in Al(IV) peak high for BC 500 A 0.5 may result from Al adopting a tetrahedral site. The NH 3 -TPD results indicate that the adsorbents contain Al Lewis-acid sites (three-coordinate Al), consistent with literature reports [37], although this is not apparent from 27 Al NMR spectra. Table 4 shows the specific surface area, pore surface area, pore volume, and zeta potential for all adsorbents. The BC 500 sample shows little change in these properties. However, the samples treated with acid, either in a single step or in combination with heat treatment (BA 0.5 , BA 0.5 C 500 and BC 500 A 0.5 ), see increases in specific surface area and pore volume; these increases may result from larger cations such as Na + , Mg 2+ , K + , Ca 2+ and Fe 3+ , including Al 3+ being replaced by very small H + cations based on the results of XRF (see Table 1) and FT-IR spectra (see Figure 2). Moreover, the zeta potential characterises the surface property of clay particles in suspension. The magnitude of this parameter is often used as a measure of the strength of the repulsive interactions between similarly charged particles. Table 4 shows the zeta potential of bentonite and modified bentonite suspension. In pH 7, the zeta potential is negative for B, BC 500 , BC 500 A 0.5 , and BA 0.5 and positive for BA 0.5 C 500 . Nonetheless, the permanent negative charge is very high in bentonite. This is due to the high degree of isomorphic substitution, where the Al(III) elements in the octahedral sheet are substituted by Fe(II) or Mg(II) in addition to some Si(IV) in the tetrahedral layer being substituted by Al(III) [40].    3.1.8. Particle Size Analysis

Textural Properties Study
The particle size distributions are shown in the supplementary information in Figure S2. The mean particle sizes of B, BC 500 , BC 500 A 0.5 , BA 0.5 and BA 0.5 C 500 are 7.15, 11.50, 8.71, 27.36, and 97.08 µm, respectively (as mentioned in the supplementary information in Table S1). The samples were treated first with acid, BA 0.5 C 500 and BA 0.5 showing bimodal particle-size distribution. The BA 0.5 sample has the maxima at approximately 11 µm (74.73%) and 25 µm (25.27%) ( Figure S2b). The large particle size seen for BA 0.5 may be due to the formation of a stable aggregate via hydrogen bonding between silanol groups, while BA 0.5 C 500 exhibits the maxima at approximately 13 µm (58.39%) and 229 µm (41.61%) ( Figure S2c). The larger particle size may result from the occurrence of dehydroxylation between the aggregated particles due to calcination. On the other hand, the particle size distributions for BC 500 A 0.5 and BC 500 show only monomodal distributions, with maxima at about 9 and 11 µm, respectively ( Figure S2d,e). The BC 500 A 0.5 particle size is smaller than that for BC 500 due to acid erosion.

The Effect of pH
The changes in pH of the solution affect the speciation of adsorbate species, adsorbent surface charge, and degree of pesticide ionisation [41,42]. Bentonite can adsorb pesticides through two different mechanisms including cation exchange and adsorption [43]. Both mechanisms are pH dependent because most silanol and aluminol groups are protonated on the surface of bentonite under acid conditions. Thus, bentonite surface exists in a typically positive charge at low pH and the charge decreases, eventually becoming negative as pH increases. Figure 6a shows the effect of pH on cationic pesticide (paraquat; PQ 2+ ) adsorption by bentonite and modified bentonite. PQ 2+ attaches to the active site through ion exchange, emerging mostly at the interlayer space of bentonite and the modified bentonite samples due to its small size, unlike the ion exchange of protonated atrazine (acidic form) that occurs very well at the face and edge of the adsorbents because of the large size of atrazine. The high adsorption of PQ 2+ appears on bentonite and BC 500 as shown in Figure 6a. The lowest adsorption of PQ 2+ is on BA 0.5 C 500 . The high amount of adsorbed PQ 2+ corresponds to the high value of CEC and negative zeta potential (see Tables 1 and 4). All adsorbents display the adsorption capacity of PQ 2+ , which increases with an increase in the pH of the solution, as shown in Figure 6a. The adsorption value of PQ 2+ increases drastically at high pH because the surface of the adsorbents becomes more negatively charged, possibly due to the SiOH − group on the surface changing to SiO − that prefers to adsorb PQ 2+ . Considering the case of atrazine adsorption by bentonite and modified bentonite shown in Figure 6c, atrazine is possibly in either protonated or neutral form depending on the solution pH. At pH values approaching a pKa of atrazine (pKa = 1.7), both protonated and neutral forms are present in the solution. Protonated atrazine preferentially adsorbs by cation exchange, while un-protonated form tends to adsorb on the active sites. According to the results of the adsorption in the range of pH 4-8, the neutral form of atrazine seems to be the most preferential to BC500A0.5. At pH 2.0, BC500 exhibits the greatest atrazine adsorption capacity. One explanation is that at pH 2, where atrazine exists in a neutral and protonated form, and the removal of atrazine would be through ion exchange and the adsorption mechanism. BC500 removes atrazine mainly through ion exchange due to its low surface area. Nonetheless, it has a lower CEC than bentonite, suggesting that the ion exchange process of cationic atrazine probably occurs efficiently at the edge and face of BC500 instead of at the interlayer spacing (like bentonite), in which a large molecule of atrazine is difficult to penetrate, causing very low adsorption for bentonite. The result of calcination might cause more exchangeable cations moving to the edge and face of BC500. Consequently, the cation exchange of atrazine is easier at these sites. In the case of adsorbents activated with acid (BA0.5C500, BA0.5, and BC500A0.5), some partial cations are removed. Thus, these adsorbents show lower CEC values, causing a reduction in ion exchange. However, these adsorbents contain high specific surface areas and a feature of Al Lewis-acid site, Si-OH and Si-OH2 + groups, which are available for binding to both forms of atrazine. At pH 2, however, the adsorbent of BC500A0.5 still shows a high adsorption capacity in spite of having a smaller specific surface area ( Table 5), suggesting that atrazine binds to BC500A0.5 through both ion exchange and the adsorption process. This may result from a small particle size with highly active sites at the edge and face of BC500A0.5 which enhances the removal of atrazine. The feasible interaction of atrazine bound to the adsorbents is a formation of a Lewis acid-base complex from an Al Lewis-acid site and Lewis base from atrazine (lone pair of electron on nitrogen), including the formation of hydrogen bonds to Si-OH and Si-OH2 + groups.
The adsorptions of diuron are shown in Figure 6d. Diuron exists half in neutral form and half in cationic form at approximately pH 4 [45]. The diuron adsorption is high for all adsorbents at low pH arising through the ion exchange process and the adsorption decreases with increasing pH, possibly due to competition between excess OH − and adsorbate to interact with active sites. However, the high adsorption of diuron with BC500 and In contrast to the case of 2,4-D adsorption (Figure 6b), 2,4-D has a pKa of 3.55; it can exist in both neutral and anionic form. At a pH of 3-6, an anionic form appears at around 22-99.5% in aqueous solution [44]. 2,4-D adsorption by BA 0.5 C 500 is found to be the highest in the pH range of 2-6, due to the positively charged surface of BA 0.5 C 500 (as in Table 4), preferring to absorb the anion form and including BA 0.5 C 500 with the highest surface area. Thus, BA 0.5 C 500 is an adsorbent suitable for 2,4-D. The lower adsorption capacity of 2,4-D with an increase in pH is probably due to the increased negative charge of the adsorbents.
Considering the case of atrazine adsorption by bentonite and modified bentonite shown in Figure 6c, atrazine is possibly in either protonated or neutral form depending on the solution pH. At pH values approaching a pKa of atrazine (pKa = 1.7), both protonated and neutral forms are present in the solution. Protonated atrazine preferentially adsorbs by cation exchange, while un-protonated form tends to adsorb on the active sites. According to the results of the adsorption in the range of pH 4-8, the neutral form of atrazine seems to be the most preferential to BC 500 A 0.5 . At pH 2.0, BC 500 exhibits the greatest atrazine adsorption capacity. One explanation is that at pH 2, where atrazine exists in a neutral and protonated form, and the removal of atrazine would be through ion exchange and the adsorption mechanism. BC 500 removes atrazine mainly through ion exchange due to its low surface area. Nonetheless, it has a lower CEC than bentonite, suggesting that the ion exchange process of cationic atrazine probably occurs efficiently at the edge and face of BC 500 instead of at the interlayer spacing (like bentonite), in which a large molecule of atrazine is difficult to penetrate, causing very low adsorption for bentonite. The result of calcination might cause more exchangeable cations moving to the edge and face of BC 500 . Consequently, the cation exchange of atrazine is easier at these sites. In the case of adsorbents activated with acid (BA 0.5 C 500 , BA 0 . 5, and BC 500 A 0.5 ), some partial cations are removed. Thus, these adsorbents show lower CEC values, causing a reduction in ion exchange.
However, these adsorbents contain high specific surface areas and a feature of Al Lewis-acid site, Si-OH and Si-OH 2 + groups, which are available for binding to both forms of atrazine. At pH 2, however, the adsorbent of BC 500 A 0.5 still shows a high adsorption capacity in spite of having a smaller specific surface area ( Table 5), suggesting that atrazine binds to BC500A0.5 through both ion exchange and the adsorption process. This may result from a small particle size with highly active sites at the edge and face of BC 500 A 0.5 which enhances the removal of atrazine. The feasible interaction of atrazine bound to the adsorbents is a formation of a Lewis acid-base complex from an Al Lewis-acid site and Lewis base from atrazine (lone pair of electron on nitrogen), including the formation of hydrogen bonds to Si-OH and Si-OH 2 + groups. The adsorptions of diuron are shown in Figure 6d. Diuron exists half in neutral form and half in cationic form at approximately pH 4 [45]. The diuron adsorption is high for all adsorbents at low pH arising through the ion exchange process and the adsorption decreases with increasing pH, possibly due to competition between excess OH − and adsorbate to interact with active sites. However, the high adsorption of diuron with BC 500 and BA 0.5 C 500 compared to the others may result from an increase in hydrophobicity (as mentioned in TG results) for both adsorbents to adsorb a more neutral form.

Simultaneous Adsorption of Various Pesticides in Solutions
The results of each pesticide adsorption (single-component) and simultaneous adsorptions of the four pesticides (quaternary-component) from aqueous solution onto all adsorbents at ambient temperature and at initial pH of solution (6.5 ± 0.5) without pH adjustment are shown in Figures 7 and 8, respectively. All adsorbents can remove PQ 2+ from both single and quaternary-component systems. In the quaternary-component system, the adsorption of PQ 2+ was slightly reduced compared with single-component due to the PQ 2+ adsorption occurring mainly inside the interlayer space of the adsorbents. Accordingly, the other pesticides (atrazine diuron and 2,4-D) in the quaternary-component do not affect the adsorption of PQ 2+ in the interlayer space of the adsorbents. The 2,4-D adsorption is decreased, except for BA 0.5 C 500 with the same adsorption capacity as in a single component, possibly due to the positive zeta potential of BA 0.5 C 500 surface (see Table 4) which provides an electrostatic attraction from the positively charged surface and 2,4-D (anion pesticide). Atrazine still prefers BC 500 A 0.5 in quaternary-component, but the adsorption capacity is lower than that in a single-component due to the limited number of active sites (as mentioned in Topic 2.6). In contrast to diuron (non-polar pesticide), it was found that the adsorption of diuron in the quaternary-component was higher than that in the single-component. This could be explained by the simultaneous adsorption; PQ 2+ adsorbed at the surface of the adsorbents can act as an additional active site, which is competent to attach diuron through the interaction between dipole (C = O of diuron) and cation (PQ 2+ ). Therefore, the presence of PQ 2+ in quaternary-component enhances diuron adsorption. The maximum total values of the simultaneous adsorption of various pesticides from aqueous solution belong to BC 500 A 0.5 , indicating that BC 500 A 0.5 is the most suitable for the simultaneous removal of different types of pesticides. Hence, it could be regarded as an effectively multifunctional adsorbent. 4) which provides an electrostatic attraction from the positively charged surface and 2,4-D (anion pesticide). Atrazine still prefers BC500A0.5 in quaternary-component, but the adsorption capacity is lower than that in a single-component due to the limited number of active sites (as mentioned in Topic 2.6). In contrast to diuron (non-polar pesticide), it was found that the adsorption of diuron in the quaternary-component was higher than that in the single-component. This could be explained by the simultaneous adsorption; PQ 2+ adsorbed at the surface of the adsorbents can act as an additional active site, which is competent to attach diuron through the interaction between dipole (C = O of diuron) and cation (PQ 2+ ). Therefore, the presence of PQ 2+ in quaternary-component enhances diuron adsorption. The maximum total values of the simultaneous adsorption of various pesticides from aqueous solution belong to BC500A0.5, indicating that BC500A0.5 is the most suitable for the simultaneous removal of different types of pesticides. Hence, it could be regarded as an effectively multifunctional adsorbent.

Adsorption Isotherms
Pesticide adsorption data were evaluated with Langmuir and Freundlich isotherm models. The Langmuir isotherm was used to determine the monolayer and homogeneous adsorption on a surface [33]. The Freundlich isotherm is an empirical equation that assumes the adsorption surface is heterogeneous during the course of the adsorption process or multilayer sorption [46].

Adsorption Isotherms
Pesticide adsorption data were evaluated with Langmuir and Freundlich isotherm models. The Langmuir isotherm was used to determine the monolayer and homogeneous adsorption on a surface [33]. The Freundlich isotherm is an empirical equation that assumes the adsorption surface is heterogeneous during the course of the adsorption process or multilayer sorption [46].
Linear forms of the Langmuir and Freundlich adsorption isotherms shown in Equations (2) and (3), respectively, were applied to quantitatively evaluate sorption performance: C e q e = C e q m + 1 q m K L (2) where C e is the equilibrium concentration of the pesticide in a bulk solution (mg L −1 ), q e is the amount of pesticide adsorbed per unit mass of adsorbent at equilibrium (mg g −1 ), q m is the maximum adsorption capacity (mg g −1 ), and K L is the Langmuir constant (L mg −1 ): where K F (units: (mg g −1 )(L mg −1 ) 1/n ) is the Freundlich constant, a characteristic of the system. The constant is indicative of the relative adsorption capacity of adsorption, while the term 1/n represents adsorption intensity. The detailed parameters of these different isotherm equations are also listed in Tables 5-8. All correlation coefficients (R 2 ) exceed 0.9, suggesting that all models fit the experimental results well.
The atrazine adsorption at pH 6 fits the Langmuir model better than the Freundlich model and the calculated maximum monolayer (q m ) of adsorbed atrazine on BC 500 A 0.5 is higher than the other adsorbents. Moreover, neutral diuron adsorption at pH 6 shows the best fit to the Langmuir model compared to the Freundlich isotherm model for BC 500 , BA 0.5 C 500, and BC 500 A 0.5 . Moreover, diuron at pH 6 was preferentially adsorbed on bentonite and the modified bentonite surfaces via monolayer adsorption. In the case of bentonite and BA 0.5 for diuron adsorption at pH 6, it is fit both for the Langmuir and the Freundlich isotherm models. In the case of anion and cation, pesticides would be described with 2,4-D and paraquat pesticide, respectively. The data for 2,4-D adsorption at pH 6 the adsorption is only anion 2,4-D. The R 2 value confirms the adsorption equilibrium data fitted well with the Langmuir model. The q m of 2,4-D adsorbed on BA 0.5 C 500 is higher than that on the other adsorbents in this experiment. Moreover, the adsorption isotherm for the removal of paraquat from aqueous solution, the R 2 values from the Langmuir isotherm of all adsorbents are close to 1, indicating that the surface of each adsorbent is unique for paraquat adsorption in monolayer adsorption.
We selected the adsorbent that is capable of adsorbing the maximum amount of each pesticide to study the adsorption isotherms. The adsorption isotherms of bentonite and the selected modified bentonite at 303, 313, and 323 K are shown in the supplementary information in Figure S3 and Tables 9-12. The results indicate that atrazine, diuron and 2,4-D adsorptions increase with higher temperatures, being the cause of increased pesticide diffusion rates onto the active sites of the adsorbents, but with little significant effect for paraquat adsorption. The maximum adsorption monolayer capacity (q m ) and the Langmuir constant (K L ), calculated from the slope and intercept of the linear plot, are shown in Tables 9-12. Enthalpy (∆H), entropy (∆S) and free energy (∆G) changes explain the thermodynamic data for the adsorption process, which were calculated by where R is the gas constant (8.314 J mol −1 k −1 ); T is temperature (K) and K L is adsorption constant in the Langmuir isotherm. The values of ∆H • and ∆S • can be calculated from the slope and intercept of the linear variation of ln K L with reciprocal of temperature (1/T) as in the following equation: The thermodynamic parameters

Kinetic Experiments
Kinetics studies on the adsorption of pesticide onto bentonite and modified bentonites were performed. The results are shown in Figure 9. Measurements were made using 20 mg adsorbent at 30 • C, with atrazine concentrations of 50 mg L −1 . All adsorbents exhibited a high affinity for atrazine during the first 50 min, and equilibrium was reached within 200 min.
11, x FOR PEER REVIEW 19 of 23 that adsorption follows the pseudo-second-order model. The indication is that the adsorption of pesticides occurs by a chemisorption mechanism between pesticide and adsorbent by innersphere complexes. The Langmuir maximum adsorption capacity for the pesticides obtained from this work was compared with other reports from the literature. Table 13 confirms that the efficiency of bentonite and modified bentonite is similar to that of the other adsorbents, but the adsorbent of BC500A0.5 is the best for atrazine removal due to the smallest particle size and the highly active region of Al Lewis-acid site which is available for easily bound atrazine via a Lewis acid-base complex. Therefore, the fixation of atrazine in BC500A0.5 is more promoted than zeoliteX, biochars (SBB) and iron nanoparticles (INPs). While ion exchange is a primary process in removing cationic pesticide (paraquat), bentonite is therefore more efficient than diatomaceous earth. For diuron adsorption, BC500 is considered a preferred diuron for hydrophobic interaction that BC500 provides high hydrophobic space. Consequently, it shows a similar result to using activated carbon.  Pseudo-first and pseudo-second-order models were employed to correlate the kinetics data. A pseudo-first-order kinetics adsorption model was suggested for the sorption of solid/liquid systems and is expressed in an integrated and linear form in Equation (6) [47]: log q e − q t = log q e − k 1 2.303 t (6) where k 1 is the adsorption rate constant (min −1 ), q e and q t are the atrazine adsorption loadings (mg g −1 ) at equilibrium and at time t (min), respectively. If the pseudo-first-order kinetics is applicable, a plot of log (q e − q t ) versus t should provide a straight line, from which k 1 and the predicted value for q e can be determined from the slope and intercept of the plot. The pseudo-second-order kinetics model is expressed by Equation (7) [48]: where k 2 (g mg −1 min −1 ) is the pseudo-second-order adsorption rate constant. A plot of t/q t against ( Figure 9) yields a straight line, from which the rate constant k 2 and q e can be calculated. The pseudo-second-order model is based on the assumption that the rate-limiting step may be chemisorption, which involves valence forces by electron sharing or electron exchange between the adsorbent and the adsorbate. Tables S6-S9 in the supplementary information shows maximum adsorption capacities of q e , k 1 , k 2 and the correlation coefficient R 2 , calculated from the pseudo-first and pseudo-second-order models. The maximum adsorption capacity, q e , calculated from the pseudo-second-order equation, is in accordance with the experimentally determined values. The R 2 values for the pseudo-second-order are greater than those in the pseudo-first-order model. These results suggest that adsorption follows the pseudo-second-order model. The indication is that the adsorption of pesticides occurs by a chemisorption mechanism between pesticide and adsorbent by innersphere complexes.
The Langmuir maximum adsorption capacity for the pesticides obtained from this work was compared with other reports from the literature. Table 13 confirms that the efficiency of bentonite and modified bentonite is similar to that of the other adsorbents, but the adsorbent of BC 500 A 0.5 is the best for atrazine removal due to the smallest particle size and the highly active region of Al Lewis-acid site which is available for easily bound atrazine via a Lewis acid-base complex. Therefore, the fixation of atrazine in BC 500 A 0.5 is more promoted than zeoliteX, biochars (SBB) and iron nanoparticles (INPs). While ion exchange is a primary process in removing cationic pesticide (paraquat), bentonite is therefore more efficient than diatomaceous earth. For diuron adsorption, BC 500 is considered a preferred diuron for hydrophobic interaction that BC 500 provides high hydrophobic space. Consequently, it shows a similar result to using activated carbon.

Conclusions
Bentonite was modified by either acid, heat, or both for the removal of different types of pesticides from an aqueous solution. It was found that BC 500 A 0.5 is the best and specific to atrazine adsorption with a very large amount compared to other adsorbents as a result of its small particle size with high edges and faces containing a high number of active sites (Al Lewis-acid sites, Si-OH, and Si-OH 2 + groups). It is remarkable that BC 500 A 0.5 is highly specific in the adsorption of atrazine mainly due to participation and strong interaction between a lone pair electron on the nitrogen of the pesticide to Al Lewis-acid site of the adsorbent and additional moderate interaction is hydrogen bond formation of amine groups to Si-OH and Si-OH 2 + groups. BC 500 shows higher diuron adsorption than the others adsorbent due to its high hydrophobicity in siloxanes; consequently, the interaction of BC 500 and diuron maybe mainly hydrophobic. Additionally, BA 0.5 C 500 is suitable for 2,4-D. The 2,4-D-adsorption may be induced by the positive zeta potential compensation indicating an interaction of anionic 2,4-D and the positively charged surface of the adsorbent.
Finally, it seems apparent that high paraquat adsorption capacity was found for all adsorbents. However, bentonite is the most favourite for paraquat as a result of the main removal of paraquat through ion exchange which is easily occurring due to a small size of paraquat and bentonite presents a high value of CEC, whereas BA 0.5 C 500 showed the lowest capacity because of BA 0.5 C 500 with low CEC and swelling. BC 500 A 0.5 was recently developed by combining steps of bentonite for simultaneous pesticide removal from an aqueous solution, because it could remove a higher total amount of pesticides than the other adsorbents, indicating that it could be used as a good multifunctional adsorbent. The adsorption of all pesticides fits well with the Langmuir adsorption model compared to the Freundlich model, and the adsorption kinetics follow a pseudo-second order (chemisorption). The thermodynamic parameters of the pesticide adsorption, such as Gibbs free energy (∆G • ), enthalpy (∆H • ) and entropy (∆S • ) were also calculated, and the parameter values are reported in spontaneous and endothermic processes. This research demonstrated that the treatment of bentonite with a sequence of the combination of heat and acid provides a high potential adsorbent for pesticide removal. It also offers a low cost and convenient method when compared with other technics.
Based on the use of artificially prepared pesticide solution in this work the effect of ionic strength on the adsorption was neglected. However, in future work, adsorptions in wastewater effect of ionic strength will be considered.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/app11115147/s1, Figure S1: TGA/DTG for bentonite (a), BA 0.5 (b), BC 500 (c), BA 0.5 C 500 (d) and BC 500 A 0.5 (e); Figure S2: Particle size distributions for bentonite (a), BA 0.5 (b), BA 0.5 C 500 (c), BC 500 (d), and BC 500 A 0.5 (e). Figure S3: Adsorption isotherm of pesticide on bentonite and modified bentonite at 303, 313 and 323 K. Table S1: Measurement of particle size distribution. Table S2: Thermodynamic parameters for adsorption of atrazine on bentonite and BC 500 A 0.5 . Table S3: Thermodynamic parameters for adsorption of diuron on bentonite and BC 500 . Table S4: Thermodynamic parameters for adsorption of 2,4-D on bentonite and BA 0.5 C 500 . Table S5: Thermodynamic parameters for adsorption of paraquat on bentonite and BC 500 . Table S6: Kinetic model parameters obtained for the adsorption of atrazine onto bentonite and modified bentonite adsorbents. Table S7: Kinetic model parameters obtained for the adsorption of diuron onto bentonite and modified bentonite adsorbents. Table S8: Kinetic model parameters obtained for the adsorption of 2,4-D onto bentonite and modified bentonite adsorbents. Table S9: Kinetic model parameters obtained for the adsorption of paraquat onto bentonite and modified bentonite adsorbents.
Funding: This work was supported by research grants to C.P. from the Royal Thai government Science and Technology Ministry Scholarship, National Science and Technology Development Agency (NSTDA).