Adsorption and Separation of Crystal Violet, Cerium(III) and Lead(II) by Means of a Multi-Step Strategy Based on K10-Montmorillonite

: A multi-step procedure, based on the employment of K10-Montmorillonite, is proposed for the selective removal of metal ions and dyes from a multicomponent solution. The objective is twofold: decontaminate the e ﬄ uents and separate and recover the valuable byproducts present in wastewaters. Three common contaminants, i.e., crystal violet dye (CV), Ce(III) and Pb(II) were chosen as “model” pollutants. The main factors a ﬀ ecting the pollutants’ sorption were investigated. The experimental data were correlated with adsorption isotherms and kinetic models to obtain a deeper insight into the adsorption processes. The a ﬃ nity of the clay toward the pollutants is favored by an increasing pH and follows the order CV > Pb(II) > Ce(III). Whereas Ce(III) metal ions do not adsorb onto clay under strongly acidic conditions, both Pb(II) and CV can adsorb under all the investigated pH conditions. The analysis of isotherms and kinetic proﬁles revealed that CV adsorbs onto clay through a mechanism consisting of two parallel processes, namely cation exchange on the external mineral surface and in the interlayer and surface complexation at the edge sites, while metal ion uptake is due solely to cation exchange processes involving mineral surfaces. The time required for the complete removal of pollutants follows the order CV > Ce(III) >> Pb(II). The possibility to modulate the adsorption features by changing experimental conditions was successfully employed to propose the best strategy for the progressive removal of di ﬀ erent components from aqueous solutions.


Introduction
Water effluents are very complex systems that usually contain different kinds of pollutants. The modern approach to water remediation requires the development of "green" and economically sustainable strategies which allow for the selective removal of the contaminants and provide the possibility of recovering and reusing them [1][2][3][4][5].
The choice of montmorillonite as sorbent was related to the extensive literature dealing with the investigation and exploitation of this mineral clay for environmental applications [34][35][36][37][38]. The present Table 1. Molar adsorption coefficient values (ε/M −1 ·cm −1 ) of crystal violet dye (CV), Pb(II) and Ce(III) at pH 2.0 and 6.0. Aiming at evaluating the minimum amount of clay required for the complete removal of the pollutants under the different conditions, experiments were performed by keeping the concentration of contaminants constant (20 mg·dm −3 and 60 mg·dm −3 ) and changing the amount of mineral clay in the range 0.05-1.6 g·dm −3 . The effect of pH was investigated by carrying out experiments in the pH range from 2.0 to 8.0. When necessary, the supernatants of Pb(II) and CV were opportunely diluted, while, for Ce(III) solutions, due to the very low ε value, long path cuvettes (optical path lengths 5 cm) were employed.

Molar Adsorption
In order to construct the adsorption isotherms, the concentration of the pollutants was changed in the range 2-300 mg·dm −3 at a constant amount of K10-Mt (0.4 g·dm −3 ).
Kinetic adsorption profiles were studied under constant stirred conditions at 25.0 • C by monitoring the UV-Vis spectrum evolution of aqueous solutions of pollutants in the presence of K10-Mt by using an Analytic Jena S600 diode array spectrophotometer. The initial concentration of the contaminants was chosen in order to obtain reasonable absorbance values (10 mg·dm −3 for CV and Pb(II) and 50 mg·dm −3 for Ce(III)). In all cases, the amount of clay was chosen in order to ensure the total removal of them. Since it is known from the literature [52] that the coalescence of the montmorillonite particles could slightly influence the absorbance reduction, this process was also monitored spectrophotometrically and automatically subtracted point by point from the kinetic adsorption profiles.

Effect of Experimental Parameters: pH Value, Initial Concentration of Adsorbate, Amount of Sorbent
Batch experiments were performed to evaluate the influence of various experimental parameters onto sorption processes. Figures 1 and 2 report the adsorption capacity values of the pollutants as a function of the amount of clay in terms of both the amount of adsorbed pollutant per gram of clay (Qe) and the percentage of sorption. Two different values of pH (namely pH 2.0 and pH 6.0) and two different values of sorbate initial concentration were analyzed.
These preliminary experiments allow for the determination of the amount of clay required for the complete removal of the different pollutants (100% sorption) under different experimental conditions.
As clearly evidenced in the figures, the equilibrium sorption capacity, Qe, which actually represents the pollutant/clay ratio, ensuring the complete adsorption of the species, follows the order CV > Pb(II) > Ce(III). As for the Ce(III), as already observed in [39], under strongly acidic conditions no adsorption Minerals 2020, 10, 466 4 of 16 onto K10-Mt is detected. Moreover, an increase in the sorption capacity with the increasing pH is detected, as is better evidenced in Figure 3.     As clearly evidenced in the figures, the equilibrium sorption capacity, Qe, which actually represents the pollutant/clay ratio, ensuring the complete adsorption of the species, follows the order CV > Pb(II) > Ce(III). As for the Ce(III), as already observed in [39], under strongly acidic conditions no adsorption onto K10-Mt is detected. Moreover, an increase in the sorption capacity with the increasing pH is detected, as is better evidenced in Figure 3. The increase in the adsorption capacity of cationic species with the increasing pH is in line with the literature [36,39] and is attributable to the deprotonation of hydroxyls, which produces a larger negative charge to compensate. The lack of adsorption of Ce(III) at an acidic pH seems to indicate the strong influence of electrostatic repulsions between H + ions and the metal cations.

Adsorption Isotherms
A deeper understanding of the adsorption processes can be achieved through an investigation of the adsorption isotherms, as reported in Figure 4. The increase in the adsorption capacity of cationic species with the increasing pH is in line with the literature [36,39] and is attributable to the deprotonation of hydroxyls, which produces a larger negative charge to compensate. The lack of adsorption of Ce(III) at an acidic pH seems to indicate the strong influence of electrostatic repulsions between H + ions and the metal cations.

Adsorption Isotherms
A deeper understanding of the adsorption processes can be achieved through an investigation of the adsorption isotherms, as reported in Figure 4.
The sorption parameters obtained from the application of the classical Langmuir and Freundlich models are collected in Table 2.
The Akaike information criteria (AIC), the coefficient of multiple determination (R 2 ) and the analysis of the residuals (Supplementary Materials, Figure S1) were used as statistical criteria to determine the most reliable model [53]. The most reliable model was chosen on the basis of the lower AIC value, the R 2 value closest to unity and the higher randomness of the results in the residual plot.
According to Parisi et al. [39], the Freundlich model, which implies the heterogeneity of the adsorption sites, proved to be the most reliable for the uptake of CV. The Langmuir model better reproduces the adsorption isotherm of both Pb(II) and Ce(III), indicating the presence of a single adsorption site for the metal ion uptake. It can be argued therefore that both cation exchange processes and electrostatic interaction with the clay surface occur in the case of CV adsorption, while the second process is solely responsible for the metal ion uptake.
Although the adsorption parameters obtained for dye and metals do not allow for a direct comparison, Figure 4 and data reported in Table 2 confirm the already-reported trend for the adsorption capacity (CV > Pb(II) > Ce(III)). The sorption parameters obtained from the application of the classical Langmuir and Freundlich models are collected in Table 2.

of 16
Additional information about the adsorption mechanism was obtained through the application of the Dubinin-Radushkevich (DR) equation ( Figure 5): where ε = RT ln(1 + 1/Ce), R (KJ·mol −1 K −1 ) the gas constant, T(K) the absolute temperature, Xm (mol·g −1 ) the clay adsorption capacity and k (mol 2 ·KJ −2 ) the DR isotherm constant related to the adsorption energy through the equation: adsorption sites, proved to be the most reliable for the uptake of CV. The Langmuir model better reproduces the adsorption isotherm of both Pb(II) and Ce(III), indicating the presence of a single adsorption site for the metal ion uptake. It can be argued therefore that both cation exchange processes and electrostatic interaction with the clay surface occur in the case of CV adsorption, while the second process is solely responsible for the metal ion uptake.
Although the adsorption parameters obtained for dye and metals do not allow for a direct comparison, Figure 4 and data reported in Table 2 confirm the already-reported trend for the adsorption capacity (CV > Pb(II) > Ce(III)).
Additional information about the adsorption mechanism was obtained through the application of the Dubinin-Radushkevich (DR) equation ( Figure 5): where ε = RT ln(1 + 1/Ce), R (KJ·mol −1 K −1 ) the gas constant, T(K) the absolute temperature, Xm (mol·g −1 ) the clay adsorption capacity and k (mol 2 ·KJ −2 ) the DR isotherm constant related to the adsorption energy through the equation:  The sorption DR parameters are reported in Table 3. Perusal of the data reported in Table 3 indicates that the obtained energy values are in the range 8-16 kJ·mol −1 , which, according to the literature [60], is characteristic of adsorption processes dominated by the cation exchange mechanism. The capability of metal ions and cationic dyes to compete with the montmorillonite counterions is well documented in the literature [61,62]. Table 3. Sorption DR parameters for the adsorption isotherms of the contaminants onto K10-Mt.

Kinetic Measurements
Kinetic adsorption profiles are reported in Figure 6 for all the investigated contaminants at two different pH values. First of all, it can be observed that the chosen values of pollutant/clay ratios ensure the achievement of the complete adsorption of the species (100% sorption).
Minerals 2020, 10, x FOR PEER REVIEW 8 of 17 The sorption DR parameters are reported in Table 3.
Perusal of the data reported in Table 3 indicates that the obtained energy values are in the range 8-16 kJ·mol −1 , which, according to the literature [60], is characteristic of adsorption processes dominated by the cation exchange mechanism. The capability of metal ions and cationic dyes to compete with the montmorillonite counterions is well documented in the literature [61,62]

Kinetic Measurements
Kinetic adsorption profiles are reported in Figure 6 for all the investigated contaminants at two different pH values. First of all, it can be observed that the chosen values of pollutant/clay ratios ensure the achievement of the complete adsorption of the species (100% sorption). First-order and double exponential kinetic models (DEM) were applied to fit the experimental data (the reader can refer to [52] and references therein for a detailed explanation of the models).
The rate equation for the first order and DEM model are described by the following equations: First-order and double exponential kinetic models (DEM) were applied to fit the experimental data (the reader can refer to [52] and references therein for a detailed explanation of the models).
The rate equation for the first order and DEM model are described by the following equations: Analogously to the procedure applied for the analysis of the equilibrium data, the discrimination among the different models was performed using the statistical criteria and the analysis of the residuals reported in the Supplementary Materials, Figure S2.
The DEM model, which describes a mechanism consisting of two parallel reactions [6,35,[54][55][56], revealed the most reliable results for the CV adsorption process, while Pb(II) and Ce(III) adsorb onto K10-Mt clay through a simple first order model. These results can be taken as further proof of the results of the adsorption isotherms, showing the presence of different adsorption sites available for Minerals 2020, 10, 466 9 of 16 CV uptake and a single adsorption site for metal ion uptake. Table 4 reports the kinetic rate constant values obtained by applying the correct models to the experimental data. Data reported in Table 4 indicate that the values of the kinetic rate constants slightly diminish with a pH increase, which is easily attributable to the higher amount of adsorbed pollutants, which means that more time is required to reach the saturation conditions. Moreover, it is worth noting that, although lead(II) and crystal violet are both able to adsorb onto K10-Mt at an acidic pH, the processes occur at very different timescales, which seems to indicate the higher affinity of the dye toward the clay in comparison with the metal ion. With the aim of finding the optimal conditions for the separation of the investigated contaminants, the kinetic profile of the adsorption onto K10-Mt at pH 2.0 of the two species simultaneously present in the solution was investigated. Figure 7 reports the spectra of a solution containing Pb(II) and CV (~10 mg·dm −3 ) treated with 0.2 g·dm −3 K10-Mt for different time values corresponding to the adsorption process.
Analogously to the procedure applied for the analysis of the equilibrium data, the discrimination among the different models was performed using the statistical criteria and the analysis of the residuals reported in the Supplementary Materials, Figure S2.
The DEM model, which describes a mechanism consisting of two parallel reactions [6,35,[54][55][56], revealed the most reliable results for the CV adsorption process, while Pb(II) and Ce(III) adsorb onto K10-Mt clay through a simple first order model. These results can be taken as further proof of the results of the adsorption isotherms, showing the presence of different adsorption sites available for CV uptake and a single adsorption site for metal ion uptake. Table 4 reports the kinetic rate constant values obtained by applying the correct models to the experimental data. Data reported in Table 4 indicate that the values of the kinetic rate constants slightly diminish with a pH increase, which is easily attributable to the higher amount of adsorbed pollutants, which means that more time is required to reach the saturation conditions. Moreover, it is worth noting that, although lead(II) and crystal violet are both able to adsorb onto K10-Mt at an acidic pH, the processes occur at very different timescales, which seems to indicate the higher affinity of the dye toward the clay in comparison with the metal ion. With the aim of finding the optimal conditions for the separation of the investigated contaminants, the kinetic profile of the adsorption onto K10-Mt at pH 2.0 of the two species simultaneously present in the solution was investigated. Figure 7 reports the spectra of a solution containing Pb(II) and CV (~10 mg·dm −3 ) treated with 0.2 g·dm −3 K10-Mt for different time values corresponding to the adsorption process.  During the first 60 min of the adsorption process, a decrease in both peaks at 591 nm and 207 nm (corresponding to CV and Pb(II), respectively) can be detected. Since the adsorption peak of Pb(II) at 207 nm is actually hidden by the broad UV-adsorption of CV, the interpretation of the experiments is not trivial. For the sake of clarity, the UV-Vis spectrum of a solution 10 mg·dm −3 (dotted curve) is also reported in the figure. A comparison with the spectrum of Pb(II) indicates that only CV adsorbs onto clay during the first 60 min of the reaction and the adsorption of the metal ion starts when CV is completely removed from the solution. The whole elimination of lead from the solution is then accomplished after 12 h.
The wavelengths at λ max for Pb(I) and CV were selected and used to construct the kinetic profiles of the pollutant's adsorption onto clay from a mixture (Figure 8). In order to eliminate the contribution of CV adsorption at 207 nm, the following procedure was applied: two experiments were performed simultaneously, i.e., adsorption of CV and lead from a mixture (A) and adsorption of CV from a solution containing just the dye (B). The kinetic profile corresponding to process B was subtracted from A, point by point, in order to delete the contribution of CV in the UV region where lead (II) adsorbs. experiments is not trivial. For the sake of clarity, the UV-Vis spectrum of a solution 10 mg·dm −3 (dotted curve) is also reported in the figure. A comparison with the spectrum of Pb(II) indicates that only CV adsorbs onto clay during the first 60 min of the reaction and the adsorption of the metal ion starts when CV is completely removed from the solution. The whole elimination of lead from the solution is then accomplished after 12 h.
The wavelengths at λmax for Pb(I) and CV were selected and used to construct the kinetic profiles of the pollutant's adsorption onto clay from a mixture (Figure 8). In order to eliminate the contribution of CV adsorption at 207 nm, the following procedure was applied: two experiments were performed simultaneously, i.e., adsorption of CV and lead from a mixture (A) and adsorption of CV from a solution containing just the dye (B). The kinetic profile corresponding to process B was subtracted from A, point by point, in order to delete the contribution of CV in the UV region where lead (II) adsorbs. The kinetic profile of lead(II) adsorption shows the presence of an "induction period" of about 60 min, corresponding to the time required for the complete removal of CV, thus indicating the higher affinity of the dye toward the clay. In other words, when simultaneously present in a mixture, the competition between the two contaminants hamper the adsorption of Pb(II).

XRD Characterization
The structural characterization of pristine and modified K10-Mt complexes was performed by means of XRD measurements (Figure 9). The kinetic profile of lead(II) adsorption shows the presence of an "induction period" of about 60 min, corresponding to the time required for the complete removal of CV, thus indicating the higher affinity of the dye toward the clay. In other words, when simultaneously present in a mixture, the competition between the two contaminants hamper the adsorption of Pb(II).

XRD Characterization
The structural characterization of pristine and modified K10-Mt complexes was performed by means of XRD measurements (Figure 9).
The XRD patterns of the clay samples showed the presence of a very weak and broad reflection d(001) at 2θ~5 • (d = 17.6 Å), characteristic of montmorillonite clay minerals, and a sharp peak at 2θ 8.9 • (d = 9.9 Å), indicating the presence of a clay mineral of the mica group (muscovite). The presence of quartz was also recognized. According to the literature [63] the weak intensity and broad peak profile of the Mt d(001) reflection is due to internal disorder, and is not necessarily indicative of low phase proportion.
A comparison between the diffraction patterns of the complexes and pristine K10-Mt indicates that the clay structure is maintained during the adsorption processes.
No changes are detected in the peak positions due to the presence of metal ions, thus indicating that Ce(III) and Pb(II) ions do not enter the clay interlayer. This is consistent with the behavior already observed by different authors [64][65][66][67], which suggests that the uptake of metal cations onto montmorillonite mainly involves the outer surface of the clay.
As for the CV pollutant, small shifts in the peak position can be observed in the diffraction pattern, thus indicating the entrance of the dye onto the clay interlayer. In order to provide a better view of these effects, in Figure 10, the superimposition of the smoothed XRD patterns in the very low angle region is reported. The XRD patterns of the clay samples showed the presence of a very weak and broad reflection d(001) at 2θ ~5° (d = 17.6 Å), characteristic of montmorillonite clay minerals, and a sharp peak at 2θ ~8.9° (d = 9.9 Å), indicating the presence of a clay mineral of the mica group (muscovite). The presence of quartz was also recognized. According to the literature [63] the weak intensity and broad peak profile of the Mt d(001) reflection is due to internal disorder, and is not necessarily indicative of low phase proportion.
A comparison between the diffraction patterns of the complexes and pristine K10-Mt indicates that the clay structure is maintained during the adsorption processes.
No changes are detected in the peak positions due to the presence of metal ions, thus indicating that Ce(III) and Pb(II) ions do not enter the clay interlayer. This is consistent with the behavior already observed by different authors [64][65][66][67], which suggests that the uptake of metal cations onto montmorillonite mainly involves the outer surface of the clay.
As for the CV pollutant, small shifts in the peak position can be observed in the diffraction pattern, thus indicating the entrance of the dye onto the clay interlayer. In order to provide a better view of these effects, in Figure 10, the superimposition of the smoothed XRD patterns in the very low angle region is reported. According to Parisi et al. [39], a slight contraction in the mica basal interlayer (d ~9.9 Å) and a small contraction in the Mt interlayer (d ~17.6 Å) are observed. The lower interplanar distance after dye adsorption could be taken as an indication of the occurrence of cation exchange processes which displace cations from the interlayer spaces.

Removal and Separation of Pollutants
The information about the equilibrium and kinetic behavior of the pollutants under the different experimental conditions was exploited to develop a procedure for the removal and separation of the different species from a solution containing the contaminants.
The proposed protocol can be resumed as follows:  According to Parisi et al. [39], a slight contraction in the mica basal interlayer (d~9.9 Å) and a small contraction in the Mt interlayer (d~17.6 Å) are observed. The lower interplanar distance after dye adsorption could be taken as an indication of the occurrence of cation exchange processes which displace cations from the interlayer spaces.

Removal and Separation of Pollutants
The information about the equilibrium and kinetic behavior of the pollutants under the different experimental conditions was exploited to develop a procedure for the removal and separation of the different species from a solution containing the contaminants.
The proposed protocol can be resumed as follows: The amount of pollutants was chosen by taking into account the extinction coefficient values of the different species (10 mg·dm −3 for CV and Pb(II) and 150 mg·dm −3 for Ce(III)) and K10-Mt was added in excess to obtain the complete removal. Figure 11 reports the spectra of the solution containing the three contaminants before (A) and after each treatment. For the sake of clarity, the spectra of the solutions of every single component are also reported. The spectrum of the supernatant reported in Figure 11B is indicative of the complete removal of the CV dye, after the treatment with K10-Mt at an acidic pH for 60 min. The successful removal of lead (II) after the subsequent treatment at an acidic pH for a longer time period (12 h) is proven by the spectrum reported in Figure 11C. Finally, the addition of K10-Mt at a higher pH value leads to the elimination of the Ce(III) metal ions.
The proposed multi-step strategy was therefore efficient for the separation and removal of pollutants of a different nature from a multicomponent solution.

Conclusions
In this study, the adsorption features of montmorillonite mineral clay were employed for the selective removal of "model" contaminants from a ternary aqueous mixture. Three widespread toxic pollutants, i.e., crystal violet, Ce(III) and Pb(II) metal ions, were chosen as representative exponents of dyes, rare earth and heavy metals.
The effects of different variables, i.e., pollutant concentration, amount of clay, pH and contact time were investigated. The adsorption processes of the single pollutants from simplified systems were analyzed by constructing the adsorption isotherms and the kinetic profiles. Significant differences in the adsorption behavior of the three species under the different experimental The spectrum of the supernatant reported in Figure 11B is indicative of the complete removal of the CV dye, after the treatment with K10-Mt at an acidic pH for 60 min. The successful removal of lead (II) after the subsequent treatment at an acidic pH for a longer time period (12 h) is proven by the spectrum reported in Figure 11C. Finally, the addition of K10-Mt at a higher pH value leads to the elimination of the Ce(III) metal ions.
The proposed multi-step strategy was therefore efficient for the separation and removal of pollutants of a different nature from a multicomponent solution.

Conclusions
In this study, the adsorption features of montmorillonite mineral clay were employed for the selective removal of "model" contaminants from a ternary aqueous mixture. Three widespread toxic pollutants, i.e., crystal violet, Ce(III) and Pb(II) metal ions, were chosen as representative exponents of dyes, rare earth and heavy metals.
The effects of different variables, i.e., pollutant concentration, amount of clay, pH and contact time were investigated. The adsorption processes of the single pollutants from simplified systems were analyzed by constructing the adsorption isotherms and the kinetic profiles. Significant differences in the adsorption behavior of the three species under the different experimental conditions were recognized. First of all, both Pb(II) and CV are able to adsorb onto clay under all the investigated pH conditions, while Ce(III) metal ions do not adsorb under strongly acidic conditions. The affinity of the clay to the pollutants is always favored by increasing pH values and follows the order CV > Pb(II) > Ce(III). The adsorption mechanism of the CV dye consists of different parallel processes, i.e., cation exchange processes and surface complexation involving both the clay interlayer and the montmorillonite surface and edges, while metal cations are held to the montmorillonite through cation exchange processes involving just the clay surface. Finally, the time required for the complete removal of pollutants follows the order CV > Ce(III) >> Pb(II).
The obtained results evidence the possibility of modulating the adsorption features by changing the experimental conditions and were employed to propose a successful approach for the progressive removal of the three different species from a multicomponent mixture. This can be viewed as a launching pad for the development of a strategy for the simultaneous decontamination and valorization of real samples of water effluents.
Supplementary Materials: The following are available online at http://www.mdpi.com/2075-163X/10/5/466/s1, Figure S1: Residuals plot for the adsorption isotherms reported in Figure 4, fitted by using the Langmuir and the Freundlich models; Figure S2: Residuals plot for the kinetic profiles reported in Figure 6 fitted by Pseudo-first order and Double Exponential Models.