Adsorption of Arsenic, Lead, Cadmium, and Chromium Ions from Aqueous Solution Using a Protonated Chabazite: Preparation, Characterization, and Removal Mechanism

The adsorption of As(V), Pb(II), Cd(II), and Cr(III) ions from aqueous solutions on natural and modified chabazite was studied. The functionalization of chabazite was performed via a protonation and calcination with the aim of generating Lewis acid sites to improve its anion exchange properties. The surface and physicochemical properties of both adsorbents were studied and compared. The adsorption isotherms of tested heavy metal ions were quantified and modeled to identify the best isotherm equation. Steric parameters for the adsorption of these ions were also calculated with a monolayer statistical physics model. Natural chabazite showed the maximum adsorption capacity for Pb(II), while the modified zeolite improved its As(V) properties in 79%. These results showed that the modified zeolite was able to remove both cations and anions from aqueous solution. The application of this functionalized chabazite can be extended for the removal of other anionic pollutants from water, thus opening the possibility of preparing new adsorbents with tailored properties for water treatment.


Introduction
Heavy metal pollution is a relevant environmental problem worldwide due to its significant potential hazard to ecosystems and human health [1].In particular, heavy metals dissolved in aqueous media are priority in terms of environmental protection because they are more toxic than the atomic form [2].These soluble heavy metals include arsenic (As), lead (Pb), cadmium (Cd), and chromium (Cr), which are nonbiodegradable and have no physiological function.They have been reported as hazardous pollutants due to their toxicological profile, high solubility, environmental persistence, and accumulation into the food chain [3,4].Therefore, these dissolved heavy metals are classified as potentially toxic elements (PTEs).
According to the International Agency for Research on Cancer (IARC), these PTEs are also considered as carcinogenic [5] where their main exposure routes are the inhalation of contaminated air, skin absorption, and ingestion of polluted food or water [6][7][8][9].Consequently, it is necessary to develop cost-effective and sustainable methods and materials to reduce the concentrations of these toxic species from polluted water and industrial effluents [10].Nowadays, the adsorption is a simple and reliable technique for water and wastewater treatment where its success mainly depends on the preparation of an effective adsorbent [11].The most used adsorbents are activated carbons, zeolites, silica gel, activated alumina, and biochar [12][13][14][15].
Natural zeolites offer several advantages for water treatment.They are low-cost minerals present in various parts of the world, have pore sizes < 20 Å, and, consequently, are capable of adsorbing or rejecting molecules, which make them suitable as molecular sieves [16].Zeolites also adjust the pH of the media, and their application usually does not generate additional environmental pollution [17].
In terms of chemical composition, zeolites are aluminosilicates with a structure consisting of a tetrahedral arrangement of silicon (Si 4+ ) and aluminum (Al 3+ ) cations with four oxygen anions (O 2-) at the vertices.This composition generates a three-dimensional framework containing SiO 4 and AlO 4 tetrahedral building blocks with permanent negative charges [18], which are balanced by monovalent or divalent counterions (e.g., alkaline or alkaline earth metals) that have the capability of acting as adsorption sites [19].The adsorption process is controlled by the zeolite properties like ion exchange capacity and selectivity [20].
Chabazite is a zeolite used worldwide due its high surface area and physicochemical properties as ion exchanger [21][22][23].Its framework has units of six double rings (D6R) interconnected through 4 rings, forming a 3-dimensional channel system that presents large ellipsoidal cavities with apertures consisting of 8 rings with a diameter of 3.8 Å in the cation-free condition [24,25].
On the other hand, the chabazite can be tailored via different processes to improve its adsorption performance for the removal of organic molecules and negatively charged species.For example, this material can be synthetized from urban wastes like cement and used to adsorb 129 mg/g of methylene blue [33].This zeolite with the incorporation of copper was also able to adsorb NOx where the results showed that Cu-CHA catalysts prepared with a conventional wet ion exchange method outperformed the NOx conversion obtained with samples prepared via chemical vapor deposition and solid-state ion exchange [34].A chemical treatment based on hexadecyltrimethylammonium bromide (HDTMA-Br) helped to change the external surface of this zeolite from negative to positive, thus being able to adsorb 26 mg/g of Cr 2 O 2− 7 [35], 6.9 mg/g of AsO 3− 4 , and 3.3 mg/g of PO 3− 4 [36].A straightforward modification process to improve the properties of zeolites relies on the decationation using different interchange ions like NH 4 Cl, NaCl, or Cu(NO 3 ) 2 with the subsequent calcination at ≥450 °C.This approach has been used to modify the surface properties of zeolites with the aim of removing anions in aqueous solution.Some examples are the removal of fluoride using clinoptilolite with a maximum adsorption capacity of 12.3 mg/g [37], the gasoline desulfurization employing 5A and 13X zeolites with a removal of 16.7 mg/g [38], and the nitrite removal from wastewater [39].The superficial zeolite modification using NH 4 Cl and a subsequent thermal treatment have been utilized to tailor synthetic chabazites for the thermal degradation of high-density polyethylene to low molecular mass hydrocarbons [40] and the adsorption of ethane (41 mg/g) and propane (81 mg/g) [41].This material had been also employed as a catalyst in the methanol to olefins reaction (forming polyaromatic coke species) [42].However, this approach has not been applied to improve the chabazite properties for the adsorption of anionic pollutants from water.Also, its impact on the cation adsorption properties of natural zeolites has not been analyzed in detail.
The aim of this study was to develop an effective and low-cost protocol to tailor the surface properties of the chabazite for the improvement of the adsorption of anionic species.This method was based on the decationation with NH 4 Cl and protonation and subsequent calcination of natural chabazite.Results showed that this functionalization route was successful to generate Lewis acid sites (LAS) on this zeolite, thus increasing its anion adsorption properties.The performance of natural and modified zeolites was tested on the removal of As(V), Pb(II), Cd(II), and Cr(III) ions.The adsorption isotherms were correlated and analyzed with the Langmuir, Freundlich, Sips, Liu, and statistical physics models.The adsorption properties of these adsorbents were compared and correlated with their ion exchange capacity and chemical composition.The mechanisms of chabazite modification and adsorption of tested ions were analyzed.Finally, this study showed that the modified chabazite can be used to remove both anions and cations of water pollutants, thus being a promising approach for low-cost water depollution.

Methodology
2.1.Preparation of Modified Chabazite.Natural zeolite used in this study was obtained from a mineral deposit from the state of Sonora, Mexico.The natural chabazite (labelled as CH-N) was sieved to obtain a particle size < 125 μm, and these zeolite particles were utilized in the surface modification process, which was implemented according to a method adapted from Ghasemian et al. [43].This method involved the following steps: (a) Homogenization: 50 g of zeolite was mixed with 500 mL of 2 N NaCl solution, stirred at 500 rpm and room temperature for 24 h [44].Zeolite was washed several times with deionized water and dried at 100 °C for 24 h.The sample obtained from the homogenization process was represented by the label "Z-Na".(b) Decationation: the dry zeolite was treated with ammonium chloride because the exchange of Na + is better for zeolites with NH + 4 [45].12.5 g of homogenized chabazite was mixed with 250 mL of 1 N NH 4 Cl at room temperature for 24 h under constant stirring.The final zeolite was washed with hot deionized water and dried at 100 °C.The sample 2 Adsorption Science & Technology obtained from the decationation process was labelled as "Z-NH 4 + ".(c) Thermal treatment: decationated zeolite was calcined at 500 °C for 3 h using a heating rate of 5 °C/min.Then, the sample was washed with deionized water and dried at 100 °C for 24 h.Finally, this modified zeolite (labelled as "CH-MS") was characterized and used for the adsorption experiments.Flow diagram of this process is included as Supplementary Information.
2.2.Chemical Characterization of Zeolite.The identification of the crystalline phases present in the zeolite samples was carried out by X-ray diffraction (XRD) analysis.XRD patterns were recorded with a diffractometer Siemens D5000 at room temperature.Samples were scanned within 2θ angular range of 5-50 °with a step size of 0.020 °for 10 s.
The textural parameters were determined by the Brunauer-Emmett-Teller method (BET) using a surface area and pore size analyzer (Quantachrome model nova 3200e).The samples were introduced into glass cell and degassed at 200 °C for 10 h.N 2 physisorption was measured at -196.15 °C.
The surface morphology of the zeolite samples was examined by using a scanning electron microscope (SEM) JEOL, model JSM-6610LV equipped with a microanalysis system model DX-4 energy dispersed spectroscopy (EDS).CH-N and CH-MS samples were placed on a fine coat brand carbon coater for this analysis.
Infrared spectra were obtained using Fourier transform infrared spectrophotometer (Thermo Scientific Nicolet iS10 FTIR).First, CH-N and CH-MS samples were dried at 70 °C for 24 h to remove the moisture.Spectra of the samples were recorded in the region of 4000-600 cm -1 .FTIR analysis was also performed to determine the acid centers of tested adsorbents via the adsorption of pyridine.Samples of CH-N and CH-MS were subjected to vacuum and then exposed to pyridine atmosphere for 24 h.FTIR spectra of pyridine-loaded samples were recorded and analyzed.
The thermogravimetric analysis (TGA) was done using a PerkinElmer equipment (model Pyris Diamond TGA/DTG) calibrated with indium and gold samples from 25 to 660 °C and 1063 °C, respectively.0.2 mg of zeolite sample was analyzed in the temperature range of 25-700 °C with increments of 10 °C/min.
The procedure established by Corbin et al. [46] was utilized for the determination of chemical composition of CH-N and CH-MS.The samples were sieved to obtain a particle size of 0.18 mm, and then, 0.1 g was mixed with 0.2 g of LiBO 2 and 0.4 g of Li 2 B 4 O 7 .The mixtures were placed in a graphite crucible with a drop of LiBr (25%w/w) and calcined for 30 min at 900 °C in a muffle.The melted mixture was dissolved in 75 mL of HNO 3 (10%w/w), filtered and made up to 200 mL.The concentrations of dissolved metals were quantified with a Thermo Scientific Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES).Multielement solutions containing Fe, Mn, Ca, Mg, K, Si, Na, Al, Sr, Pb, Sn, Hg, Cd, Zn, As, Ni, Cr, Mo, and Ba were used to prepare the calibration curve where high purity standard solutions (1000 mg/L) of each metal and deionized water were employed for this purpose.

Surface Properties of Zeolites.
The pH of the point of zero charge (PZ) of the zeolites was determined preparing solutions with pH between 2.27 and 6.75 (with a constant ionic strength of 0.01 M) via the mixing of certain volumes of 0.01 M NaOH, HNO 3 , and NaNO 3 with 0.005 g of the samples at 25 °C under constant stirring of 30 rpm for 5 days in an orbital shaker.Finally, the zeta potential of the zeolite particles was obtained using the zetameter (Zetasizer, Malvern model Nano series).
Cation exchange capacity (CEC) was obtained using a methodology reported by Ming & Dixon [47] where 1 g of CH-N and CH-MS samples was saturated with Na + ions (using 50 mL of 1 N C 2 H 3 NaO 2 for 24 h) and washed, and the exchange of Na + ions for NH 4 + ions was done using 50 mL of 1 N C₂H₇NO₂ at room temperature for 24 h.Na + ion concentration was measured with an ICP-OES, and the CEC (meEq/g) was determined with the following equation: where ½Na + is the Na + concentration (mg/L), V is the volume (L), m is the mass of chabazite (g), and MW is the Na + molecular weight (mg/mEq).

Adsorption Studies Using
Zeolites.Adsorption studies were conducted with both zeolites CH-N and CH-MS in glass Erlenmeyer flasks containing 0.05 g of the adsorbent and 0.040 L of the pollutant solution.These studies were performed at pH 4 where the solution pH was adjusted by adding HNO 3 or NaOH and mixing with a glass stirrer.Note that this solution pH was selected to avoid the precipitation of Cd(II) and Pb(II) ions as hydroxides (i.e., Cd(OH) 2 and Pb(OH) 2 at pH > 5 and 6, respectively) [48,49] and to prevent the zeolite dealumination [31,50].Other studies have also reported that a maximum adsorption capacity of Pb(II), Cd(II), and Cr(III) ions can be obtained with different adsorbents at pH 4 [48,51,52].Initial concentrations of 1-59, 20-500, 10-112, and 5-93 mg/L for As(V), Pb(II), Cd(II), and Cr(III), respectively, were used in the adsorption experiments.These solutions were prepared with J.T. Baker standard solution of 1000 mg/L of each metal and deionized water.The suspensions (adsorbent+pollutant solution) were stirring at 50 rpm and 25 ± 0:5 °C in a thermoregulated orbital shaker (IKA model KS 4000) for 5 days.This time was enough to reach the equilibrium according to preliminary adsorption experiments at tested conditions.The quantification of metal concentrations was done using ICP-OES.The calibration curve for the quantification of these metals was obtained with solutions that were prepared using a high purity standard solution (1000 mg/L) of each adsorbate and deionized water.The removal of ions per unit mass of adsorbent (i.e., adsorption capacity) was expressed as q e (mEq/g) and calculated with the mass balance given by the following equation: where C o and C e are the initial and equilibrium concentrations (mEq/L) of tested ion in the aqueous solution.Finally, the speciation of tested heavy metals was obtained by means of Hydra-Medusa speciation software at pH 0-14 and 25 °C.
2.5.Isotherm Modeling.The experimental equilibrium data were modeled to analyze As(V), Pb(II), Cd(II), and Cr(III) adsorption and to explain the interactions of these heavy metal ions with CH-N and CH-MS samples.The best isotherm model for the removal of these pollutants was also identified.Therefore, four different isotherm models were tested using a nonlinear regression procedure.They corresponded to two-parameter (Langmuir and Freundlich) and three-parameter (Liu and Sips) isotherm models.
Langmuir is a monolayer adsorption model where the adsorbent saturation occurs assuming active sites with the same energy.This model considers that there is only one interaction corresponding to one active site per specie, and this isotherm is presented in equation ( 3) [53].
Freundlich is a multilayer adsorption model where the energy of active sites is not homogeneous, and this model is commonly applied to analyze heterogeneous systems.This isotherm model is presented in equation ( 4) [54].Alternatively, Sips model can be reduced to the Freundlich and Langmuir models at low and high concentrations, respectively, and it presents the advantages of these two models since it is a combination of them.This isotherm model is given by equation ( 5) [55].The Liu isotherm is a Langmuir-Freundlich type isotherm that has the advantage over the Sips model because its exponent has no restrictions in terms of its maximum value where it is >0.This isotherm is presented in equation ( 6) [56].In summary, the corresponding equations of these isotherm models are given by where q L , q S , and q LF are the maximum adsorption capacities of the Langmuir, Sips, and Liu equations (mEq/g); K L , K S , K LF , and K F are the constants of the Langmuir, Sips, Liu (L/mEq), and Freundlich (mEq/g)(L/mEq) 1/nF isotherms; and n F , n S , and n LF are the dimensionless exponents (parameters related to the adsorption intensity that represents the relative distribution of the heterogeneity and the energy of the adsorption sites) of the Freundlich, Sips, and Liu equations, respectively.A nonlinear least squares method based on an optimization algorithm was used to obtain the isotherm model parameters.The Solver tool of Microsoft Excel software was used to minimize the sum of the square error (ERRSQ) [57]: where q t,meas is the adsorption capacity from the experiment (mEq/g), q t,calc is the adsorption capacity calculated with the isotherm model (mEq/g), and N is number of experimental points.
To compare and assess the performance of the isotherm models, three statistical metrics were calculated: determination coefficient (R 2 ), the average percentage deviation (%D), and chi-square (X 2 ) [57][58][59].These metrics are given by the following equations: A monolayer adsorption model [60] was also utilized in the isotherm data correlation with the aim of analyzing the main steric parameters involved in the removal of these ions.This model was defined by where n ads is the number of ions adsorbed per functional group of zeolite, N zeolite (mEq/g) is the quantity of zeolite functional groups involved in the adsorption of these pollutants, and C med (mEq/L) is the adsorbate concentration at the half saturation condition, respectively.

Results and Discussion
3.1.Surface Chemistry Characterization of CH-N and CH-MS.Figure 1 shows the X-ray diffraction patterns of both zeolites used in this study.For CH-N, it was possible to confirm the presence of the expected chabazite structure characterized by its main peaks at 9.39 °(1 0 1), 20.43 °(-2 1 0), and 30.37 °(-3 1 1) according to the JCPDS card 34-137 [61].This result also agreed with the calculated XRD patterns reported for this zeolite [62].X-ray diffraction pattern of CH-MS indicated that the crystalline structure remained after the modification process, thus preserving all the intensity peaks similar to CH-N.These findings could suggest that the modification took place only on the external surface, which agreed with the previous studies [63,64].Additionally, it was possible to observe an increase in the intensity of the peaks in the modified zeolite.This observation agreed with other studies (e.g., [37]) where an increment of the intensity after a similar modification to a natural zeolite was also identified.The textural properties of CH-N and CH-MS are given in Table 1.According to these results, CH-N surface area (324.4 m 2 /g) was higher than some values reported in the literature for natural zeolites such as 2.7 m 2 /g for perlite consisting of zeolites X (FAU type), P (GIS type), and phillipsite [65]; 20.3, 42, and 258 m 2 /g for clinoptilolite [66][67][68]; 68 and 305 m 2 /g for modernite [68,69]; and 202 m 2 /g for a natural chabazite from Mexico [31].However, this value was lower than 1100 m 2 /g of a natural chabazite supplied by Minerals Research-Clarkson, NY, USA [70].
Regarding the texture properties of CH-N and CH-MS, the differences between their surface areas and pore diameters were 0.52% and 3.35%, which were not significant because they were within the error interval of the equipment used in sample analysis.Nevertheless, the difference in pore volume corresponded to an increment of 8.33% that was statistically significant.It is convenient to remark that the behavior of the texture properties of different zeolites submitted to the similar surface modification treatment has been associated to the changes on their micropore structure.Both increments and reductions in these textural parameters have been reported.For instance, Saucedo-Delgado et al. [37] reported an increment of 90.6 and 22% in the surface area and pore volume of clinoptilolite.On the other hand, Albayati and Kalash [71] proved that the functionalization (via the reaction between the silylating agent and the silanol groups) of mesoporous silica MCM-41 incremented the surface area from 28.7 to 300 m 2 /g, while the calcined MCM-41 showed the highest surface area (i.e., 1000 m 2 /g).In the contrary case, a decrease of 15.5 and 23% in pore diameter and surface area of modernite was reported by Mori et al. [72], while Atiyah et al. [73] obtained surface areas of 845 and 45 m 2 /g for the mesoporous silica SBA-15 and its functionalized form NH 2 /SBA-15, respectively, thus indicating a significant decrement of this textural parameter.
SEM micrographs of chabazites showed the typical rhombohedral crystals that were consistent with the previous studies [31,74,75].By comparing these micrographs, it was possible to observe that CH-N (Figure 2(a)) presented a well-defined structure, while CH-MS (Figure 2(b)) showed partially formed rhombohedral structures.This could be explained because during the modification process, the minerals were washed several times, which could fragment the zeolite agglomerates.Chemical analysis of CH-N crystals by SEM-EDX (Figure 2(a)) indicated the presence of the characteristic elements of zeolites: Si, Al, O, Mg, Ca, K, and Na.On the other hand, EDX results showed that the exchangeable cations (Na + and Ca + ) of CH-MS (Figure 2(b)) were expelled out from the framework, where this effect agreed with the result obtained by Barrer et al. [76] and Ghasemian et al. [43].This behavior can be explained by the fact that, during the homogenization phase carried out for chabazite, Na + displaced the majority of Ca 2+ .Then, the sodium ions were exchanged with  the ammonium ions when the zeolite was treated with NH 4 Cl because the thermodynamic affinities for the exchange reactions in chabazite followed the next sequence: NH + 4 > Na + > Ca 2+ , which were given by the standard free energies of exchange [76].
Figure 3(a) presents the FTIR spectra of CH-N and CH-MS at 4000-600 cm -1 .At 750-650 and 1250-950 cm −1 , these spectra displayed the internal vibrations of TO 4 tetrahedron (T = Si and Al) named as OH stretching (indicated by the four-membered ring structure) and asymmetry stretching.The absorption band between 570 and 635 cm -1 indicated 6membered rings.The external vibrations of the tetrahedron (i.e., T-O double ring, symmetry and asymmetry stretching) appeared at 650-500, 820-750, and 1150-1050 cm -1 , respectively [77].The absorption bands observed at 1640, 3600, and 3380 cm -1 were characteristic to the water bending vibration, isolated, and H-bonded hydroxyl groups, respectively [78].These three bands for CH-N sample showed higher intensities than those observed for CH-MS, thus indicating that CH-N had a higher water content that was lost in the modified zeolite upon calcination.Another major difference was that CH-MS sample showed an absorption band at 3640 cm −1 corresponding to the hydroxyl group formed during the modification process [79].Both types of vicinal hydroxyl groups were observed via the bands at 3640 and 3700-3740 cm -1 that was associated to the silanol nests [80].
It was possible to observe that CH-MS did not present the characteristic bands of NH + 4 form (i.e., the Lewis bond NH 3 at 1630 cm -1 , while NH + 4 is characterized by bands between 1400 and 1500 cm -1 and NH stretching frequencies at 3800 cm -1 ) because they disappeared due to the heating at 290 °C [81].
Figure 3(b) reports the FTIR spectra of CH-N and CH-MS samples loaded with pyridine at 1400-1700 cm -1 .The absorption bands at 1640 and 1545 cm -1 have been attributed to the pyridine adsorption and were characteristic of the Brönsted acid sites (BAS).The bands located at ~1430 and 1490 cm -1 corresponded to hydrogen-bonded pyridine (Hpyridine), while the bands identified at 1590 and 1450 cm -1 have been attributed to coordinately bound pyridine, thus being an indicator of Lewis acidity [82][83][84].The intensity of the absorption bands of LAS (1450 and 1590 cm -1 ) and BAS (1640 cm -1 ) in the spectrum of pyridine-loaded CH-MS were higher than those of the other zeolite sample.This was due to the modification where BAS and LAS sites were formed.Emeis [85] reported similar results where pyridine adsorption was studied for natural and calcined modernite at 550 °C for 2 h.Other studies of Ward [86] showed an infrared spectrum of pyridine adsorbed on a rare earth Y zeolite calcined at 430 and 680 °C, while Wu and Weitz [87] reported the results for a ZSM-5 zeolite modified (protonated with nitrate salt and calcined at 500 °C).°C was attributed to the humidity loss of the zeolite, while the second step at 118 °C corresponded to the loss of structural water from the zeolite.The decomposition of calcium carbonate was associated to the peak at 270 °C, which was consistent with the study of Stakebake [88] that indicated that this phenomenon could occur before 310 °C.
For the case of CH-MS sample, the loss of water adsorbed on the zeolite (humidity) occurred at 109.98 °C.It is convenient to highlight that there was no other peak cor-responding to the loss of structural water for this sample, which occurred during the zeolite calcination at 500 °C.It should be noted that other studies have reported the decomposition of NH 4 Cl at 310 °C [89].Based on the DTG curve of CH-MS, it was concluded that this compound was not present in the modified zeolite.This finding could be due to the previous release of NH 3 in the calcination stage of the modification process, which was consistent with the FTIR results that confirmed the absence of this compound.
The chemical composition of CH-N and CH-MS is presented in Table 2.It can be observed that the weight     Table 2 shows Si/Al ratio of 3.3 for the CH-N sample that matched with the value of 3.2 found by Leyva-Ramos et al. [31], which was also between the range of 1.43-4.18,3.2-3.8,and 2-4 obtained by Gottardi & Galli [90], Zamzow et al. [91], and Metwally & Attallah [92], respectively.Si/Al ratio of CH-MS sample was 2.8, which was lower than the value found for CH-N sample.The decrease in the Si/Al ratio of these samples coincided with the results reported by Panayotova [93] where a natural zeolite from Kardjali was modified by heating at 373 °C and using 2 mol/dm 3 of NaOH solution.
The significant decrease in the Si/Al ratio between CH-N and CH-MS samples was caused by the superficial modification of the zeolite, whose steps are schematized in Figure 5.After homogenization (Z-Na) and decationization (Z-NH 4 + ), NH 3 was removed by a thermal treatment at ≥290 °C.This sample was named protonated form of zeolite (H-Z).This process generated BAS and was accompanied by the formation of free protons that attacked the tetrahedral aluminum lattice, thus causing the formal bond between aluminum and oxygen broken, and the corresponding formation of hydroxyl groups as silanol (Si-OH).This process was reversible since the interaction of silanols with aluminum formed the protonated zeolite and vice versa [79,94].Note that Iler [95] established that the condensation of silanol groups was higher in the presence of excess salts at pH 4-8.Therefore, the pH 7 and the application of 2 M NH 4 Cl solution in the chabazite modification process provided the optimal conditions for the formation of silanol groups.They can be eliminated with the consecutive washing of material during the modification process.Therefore, it was reflected as an increase in the proportion of Al in the zeolite framework and a change in the Si/Al ratio.

Surface Properties of Zeolites.
Figure 6 shows the results of the pH of PZ for both zeolites.It was concluded that the CH-N and CH-MS surfaces were negatively charged at tested adsorption conditions.Similar findings have been reported in the literature for clinoptilolite and other zeolites treated with NH 4 Cl and calcination process [86,96].
The negative charge for CH-MS was an expected result, which can be explained considering that the modification process was carried out at high temperature (i.e., 500 °C).Ward [86] indicated that if the calcination process occurred at ≥480 °C, BAS became LAS because the bounds of H-Z were very weak and easily removable (i.e., this process is called dehydroxylation or water loss).This phenomenon caused the formation of trigonal silicon ions and accessible trigonal aluminum [81] that can accept a pair of electrons, and therefore, it can adsorb anions.Figure 5 illustrates the formation of LAS where the negative charge of CH-MS is clearly outlined.The surface interactions involved in the adsorption of anions on this modified zeolite can be explained as follows: CH-MS zeolite showed a negatively surface charge in the aqueous solutions that caused the attraction of H + ions from the solution, while the HO -ions were attracted by the trigonal silicon ions, thus forming silanols.The adsorbate anions in the solution can interact via electrostatic attraction with the LAS [97] since they are the active sites in the zeolites [98].This mechanism agreed with that proposed by Uytterhoeven [81] for water, Lewis acid centers, and molecules with lone pairs of electrons such as NH 3 .On the other hand, the cations can be attracted by the negative charges of the CH-MS during the adsorption process.Note that the divalent cations could interact with more than a single site of the zeolite during the interface phenomenon [99].It could be expected that the adsorption of cations on CH-MS was associated to an ion exchange.As explained, the interaction of silanol groups with acid centers of a zeolite generates a protonated zeolite.But this process was not reversible at 25 °C because the zeolite was not reheated and the dehydroxylation did not occur.
CEC of CH-N and CH-MS was 1.68 and 0.53 mEq/g.These values showed a significant decrease of 68.45% of this property with respect to a CH-N sample.This result was because the ion exchange behavior of zeolites depends on the internal properties like chemical composition, framework  [100], besides the external properties like size, shape, charge, and concentration of ions [101].As stablished by Jacobs and Leuven [102], the average reaction rate per site will be related with the chemical composition, thus generating an increment almost linearly of the frequency of silanol groups with a decrement of Si/Al ratio of the framework [103], which was coupled with the formation of LAS containing trigonal silicon ions.This explained that the change in the Si/Al ratio of CH-MS sample caused a decrement of the number of active sites because of the formation of the silanol groups and the trigonal silicon ions.This pointed out that the silanol groups were not active centers such these cannot serve as adsorption sites [80].Although LAS were active sites, they can attract Sips * q model (q L , q LF , and q S ) is the maximum adsorption capacities of the Langmuir, Liu, and Sips equations.K model (K L , K LF , K S , and K F ) is the constant of the Langmuir, Liu, Sips (L/mEq), and Freundlich (mEq/g) (L/mEq) 1/nF equations.n model (n LF , n S , and n F ) is the dimensionless exponent of the Liu, Sips, and Freundlich equations.
anions.As a result of the interaction of BAS and LAS, the zeolites present a superacidity [104] since both sites act synergistically potentiating the acidity [105].As stablished by Barthomeuf [106], an acidic environment in zeolites affects the number of interchange sites, their location, density, and efficiency.
The value of CEC of CH-N was within the range of values reported for natural chabazites from Japan, Christmas Arizona, and a zeolite provided by American Colloid Co., which showed values of 1.87, 1.95, and between 0.08 and 2.61 mEq/g, respectively [20,70,107].The decrease in CEC agreed with Ferrer [108] that used a clinoptilolite submitted to the same surface treatment and subsequent calcination at 500 °C.These authors reported CEC values of 11.76 and 6.33 mEq/g for natural and modified materials, respectively.This treatment generated a change of 46% in this property, 13 Adsorption Science & Technology which was lower than that obtained with chabazite.In this direction, Moreno-Tost et al. [109] use Cu-exchanged mordenite and clinoptilolite with the subsequent calcination at 550 °C.These authors found a change in the CEC of 86% with respect to a natural clinoptilolite and 68% for natural modernite.
3.3.Analysis and Modeling of Adsorption Isotherms.Figure 7 shows the speciation diagrams of the different metals in the aqueous solution as a function of solution pH.The predominant ionic forms for tested adsorbates at pH 4 were Pb 2+ and Cd 2+ for Pb(II) and Cd(II), while H 2 AsO − 4 and ðCrOHÞ 2+ were the dissolved species for As(V) and Cr(III) during the adsorption experiments.Therefore, the experimental isotherms of the adsorption of As(V), Pb(II), Cd(II), and Cr(III) on CH-N and CH-MS samples are reported in Figure 8, and the fitting results using the models of Langmuir, Freundlich, Sips, and Liu are shown in Table 3.
The adsorption capacity on CH-N and CH-MS followed the next sequence: Pb(II) > Cr(III) > Cd(II) > As(V), which agreed with the results obtained by Caputo and Pepe [32].Experimental data showed that As(V) adsorption capacity was 0.034 and 0.061 mEq/g for CH-N and CH-MS, respectively, thus indicating that the zeolite functionalization improved in 79% its removal performance for this ion.Pb(II) adsorption capacities of CH-N and CH-MS were 1.10 and 0.41 mEq/g, respectively, while the Cd(II) and Cr(III) adsorption capacities of these zeolites were 0.29 and 0.11 mEq/g and 1.02 and 0.69 mEq/g, respectively.These reductions (33-67%) in the adsorption capacities of these cations were directly related to CEC decreased 68.45% due to the zeolite modification and the generation of BAS and LAS on the adsorbent surface.These adsorption capacities were not attributed to the surface area and pore diameter of tested zeolites, since both samples showed similar textural parameters.
The maximum exchange degree of Pb(II) on CH-N was an expected result, which can be explained according to the Eisenman-Sherry theory [110,111].Particularly, Pb(II) ion has a lower value of hydration enthalpy and a higher value of equilibrium constant (which is generally utilized as a measure of the cation exchange selectivity) in chabazite than those reported for the other two cations.
The chabazite has four classes of cation sites, named C1, C2, C3, and C4, where C1 sites are located in the center of the small cage (D6R) and they can be accessed through 6membered ring, thus representing 20% of the overall CEC of this zeolite.C2-C4 sites are located in the large ellipsoidal cavities, and they can be accessed via 8-membered ring [112,113].According to the Double Selectivity Model (DSM) proposed by Pepe et al. [114], these types of sites are named group "I" and group "II."It could be expected that large ions like ðCrOHÞ 2+ could not access to C1, while Pb(II) and Cd(II) can interact with these sites [115,116].So, the reduction of CEC in the modified zeolite was mainly associated to the loss of active sites in group I and group II, which were involved in the generation of Lewis acid centers.This phenomenon affected the final adsorption properties of the modified zeolite for cationic species.It is also important to mention that the significant reduction (up to 67%) of Pb(II) adsorption properties of CH-MS was associated to Na + ions that were expelled out from the zeolite framework during the modification process.Previous studies have reported that Na + can be completely exchangeable by Pb 2+ , thus explaining the adsorption capacities of natural chabazite [116,117].
Table 3 indicates that the q L values obtained with the Langmuir model were lower than q LF and q S values calculated by the Liu and Sips models.These findings agreed with other studies reported for the adsorption of As(V), Pb(II), Cd(II), and Cr(III) ( [118][119][120][121]).On the other hand, the n F values of the Freundlich model ranged from 1.60 to 9.71 for the adsorption of tested ions, thus indicating a favorable removal process [54,122].The Langmuir constant K L can be associated to the bonding energy between the zeolite and ions [123], and this parameter ranged from 1.21 to 190.38 (L/mEq), thus suggesting a good adsorption affinity between the zeolite and tested ions.
In terms of modeling, X 2 has been widely used to analyze the error distributions of data sets and to identify the most suitable equation for the calculation of adsorption isotherms [58,124,125].Therefore, X 2 was employed to characterize the performance of tested isotherm models for both zeolites.This statistical metric indicated that the three-parameter isotherm models (i.e., Liu and Sips) showed the better fit of experimental data, thus presenting X 2 values lower than those obtained for the two-parameter isotherm models of Langmuir and Freundlich.Herein, it is convenient to recall that the use of the Sips equation is usually restricted to the fact that its dimensionless heterogeneity factor n S must be between 0 and 1 [55,122].Results reported in Table 3 indicated that the calculated n S values for As(V), Pb(II), Cd(II), and Cr(III) adsorption on chabazite modified did not comply with this restriction.Therefore, the Liu isotherm (X 2 = 4:76 × 10 −5 − 0:075) was selected as the best model to correlate these experimental adsorption isotherms (see Figure 8).
For illustration, Table 4 shows a comparison of the maximum adsorption capacity for the removal of different ions using several adsorbents.Overall, the results confirmed that the protonated chabazite was more effective for the removal of As(V) than other zeolites reported in the literature like clinoptilolite, modernite, and chabazite [126][127][128], as well as other waste materials like hydrated cement and bone char [118,129].However, the adsorption capacities of CH-N and CH-  [130,131].This study confirmed that the functionalization of zeolites to remove anions is a promising approach to improve the performance of these materials in water treatment.For example, the clinoptilolite modified with NH 4 Cl plus its subsequent thermal treatment at 500 °C increased its fluoride adsorption capacity by 2.33-fold with respect to the natural zeolite [37].Other study showed that a natural zeolite rock treated at 800 °C increased its anion adsorption capacity with respect to the uncalcined zeolite [39].Another example is the thermal treatment (at ≥500) of different mesoporous silica sieves like MCM-41, MCM-48, SBA-15, and Co/MCM-41 to improve the removal of sulfur from diesel [132,133].
With respect to the performance of natural chabazite for the removal of cations, its Pb(II) adsorption properties are competitive and can outperform the results obtained using watermelon rind and other zeolites like natural and modified clinoptilolite and chabazite [26,138,139].Even after its functionalization, the protonated chabazite showed a Pb(II) adsorption capacity higher than those of cocoa pod husk and natural clinoptilolite [136][137][138].However, this modified zeolite can be outperformed by, for example, carbon foam and zeolites synthesized by complex processes from lithium leach residue and bagasse fly ash ( [135,141,142]).
For the case of Cd(II) adsorption, CH-MS can outperform erionite and clinoptilolite [120], while the adsorption properties of CH-N were better than those reported for carbon nanotubes and biomass wastes like peanut husk and corn stalk [143,144,156].However, fly ash treated with NaOH outperformed this zeolite [147].Also, the adsorption capacities of natural and modified chabazite for Cd(II) removal were significantly lower than those obtained with activated carbon [148,149].
For the case of Cr(III) and Cr(VI) adsorption, CH-MS outperformed different activated carbons, magnetic zeolite/ chitosan composites, multiwalled carbon nanotube, and natural clinoptilolite [137,[150][151][152]. On the other hand, the adsorption properties of CH-N were higher than those reported for magnetic zeolite/chitosan composites and porous Fe 3 O 4 hollow microspheres/graphene oxide composite [153,154].However, its removal performance was lower than the reported for functionalized multiwalled carbon nanotube with H 2 SO 4 /HNO 3 solution and ionic liquid impregnated exfoliated graphene oxide [152,155].
In summary, these results indicated that CH-MS could be considered a promising adsorbent for the removal of anions and cations from polluted water.This adsorbent could be applied in the depollution of other matrices like soil or mining waste, which also contain both types of ionic pollutants.
With respect to the interpretation of the adsorption mechanism, the steric parameters related to the adsorption of Cd(II), Pb(II), Cr(III), and As(V) on CH-N and CH-MS are reported in Table 5.In particular, n ads ranged from 0.41 to 0.76 for the adsorption of cations and n ads = 0:69 for the anion adsorption on CH-N.These results suggested that the removal of Cd(II), Pb(II), Cr(III), and As(V) can be classified as multianchorage where two adsorption sites from this zeolite could participate during the removal process [60].For the case of modified chabazite, the calculated n ads values ranged from 0.39 to 0.89 for the adsorption of heavy metal cations, while n ads was 1.68 for the adsorption of As(V) anions (see Table 5).These results indicated that the Pb(II), Cd(II), and Cr(III) adsorption prevailed as multianchorage, and a multiionic adsorption occurred for the As(V) removal with CH-MS sample.The amount of adsorption sites of CH-N that were involved in the adsorption of these pollutants was 0.12-1.19mEq/g, while the adsorption sites of CH-MS that participated in the adsorption were 0.04-0.48mEq/g.These calculations were consistent with characterization results and CEC values that indicated the loss of some adsorption sites due to the zeolite modification protocol.As illustrated in Figure 5, the framework vacancy generated during the surface modification of chabazite allowed that its adsorption sites could interact with more than one As(V) anion, thus causing a multiionic adsorption of this pollutant.

Conclusions
This study has proved that a low-cost and straightforward protocol can be used to modify and tailor the surface properties of chabazite for the adsorption of anionic pollutants from aqueous solutions.This zeolite surface modification allowed to generate Lewis acid centers that could act as electron pair acceptors, thus increasing up to 79% the As(V) adsorption properties.However, the adsorption of Pb(II), Cd(II), and Cr(III) cations of this modified chabazite decreased 33-67%.This reduction of the cation adsorption properties was generated by a decrement of its cation exchange capacity and the loss of Na + exchange sites from zeolite surface.Statistical physics calculations indicated that the adsorption of all these ions on natural chabazite was a multianchorage phenomenon, while the As(V) adsorption on modified chabazite was a multi-ionic process.A reduction of the number of zeolite adsorption sites after surface functionalization was confirmed that affected the removal of cationic species.Natural chabazite can be considered an outstanding adsorbent for the Pb(II) removal with adsorption capacities higher than 100 mg/g.The application of this modified chabazite can be extended for the removal of other relevant anion pollutants from water.Further studies should be also focused on its application to depollute other real-life matrices (e.g., soil and wastewater) containing both anionic and cationic toxic compounds.

Figure 1 :
Figure1: X-ray diffraction patterns of CH-N and CH-MS.Label "Theoretical" corresponds to the calculated pattern from International Zeolite Association, which was used as a reference pattern.

Figure 4 :
Figure 4: Results of TGA and DTG for the zeolite samples: (a) CH-N and (b) CH-MS.

Figure 5 :
Figure5: Illustrative 1D scheme of the stages of CH-N modification process where the formation of silanol, Brönsted acid sites (BAS), Lewis acid sites (LAS), and trigonal silicon ions is shown.The interaction of the CH-MS with divalent cations M(II) and anions like As(V).This figure was adapted from Uytterhoeven et al.[81] and Juzsakova et al.[45].

Figure 6 :
Figure 6: Zeta potential versus pH for CH-N and CH-MS samples.

Figure 7 :
Figure 7: (a) Arsenic, (b) lead, (c) cadmium, and (d) chromium speciation as a function of solution pH in aqueous solution.

Figure 8 :
Figure 8: Isotherms for the adsorption of (a) As(V), (b) Pb(II), (c) Cd(II), and (d) Cr(III) on CH-N and CH-MS zeolites at pH 4 and their correlation with Liu model.

Table 1 :
Textural properties of CH-N and CH-MS samples.

Table 2 :
Chemical composition of the chabazite and its modified form.
fractions of exchangeable cations in CH-N decreased in the following order: Ca 2+ , Mg 2+ , Na + , and K + .On the other hand, the chemical composition confirmed the EDX results of CH-MS (Figure2(d)) where the exchangeable cations Na + and Ca + were expelled out from the framework.

Table 3 :
Results of isotherm fitting using the Langmuir, Freundlich, Sips, and Liu equations for the adsorption of As(V), Pb(II), Cd(II), and Cr(III) on CH-N and CH-MS zeolites.

Table 4 :
Maximum adsorption capacities for the removal of As(V), Pb(II), Cd(II), and Cr(III) using different materials.

Table 5 :
Calculated steric parameters for the adsorption of Cd(II), Pb(II), Cr(III), and As(V) on CH-N and CH-MS zeolites.