Simultaneous Removal of Heavy Metals (Cu, Cd, Cr, Ni, Zn and Pb) from Aqueous Solutions Using Thermally Treated Romanian Zeolitic Volcanic Tuff

Increased concentrations of heavy metals in the environment are of public health concern, their removal from waters receiving considerable interest. The aim of this paper was to study the simultaneous adsorption of heavy metals (Cu, Cd, Cr, Ni, Zn and Pb) from aqueous solutions using the zeolitic volcanic tuffs as adsorbents. The effect of thermal treatment temperature, particle size and initial metal concentrations on the metal ions sorption was investigated. The selectivity of used zeolite for the adsorption of studied heavy metals followed the order: Pb > Cr > Cu > Zn > Cd > Ni. The removal efficiency of the heavy metals was strongly influenced by the particle sizes, the samples with smaller particle size (0–0.05 mm) being more efficient in heavy metals removal than those with larger particle size (1–3 mm). Generally, no relevant changes were observed in heavy metals removal efficiency for the treatment temperatures of 200 °C and 350 °C. Moreover, at a higher temperature (550 °C), a decrease in the removal efficiencies was observed. The Cd, Zn, Cu, Cr, Zn and Ni sorption was best described by Langmuir model according to the high values of correlation coefficient. The pseudo-first-order kinetic model presented the best correlation of the experimental data.


Introduction
In recent years, there have been increasing public health and environmental concerns related to the contamination with heavy metals [1]. The main sources of these elements in the environment comprise mostly mining, smelters, and other metal-based industrial operations [2]. Metals such as Cd and Pb have no known biological functions and are considered non-essential elements, being toxic even at very low concentrations. Cd affects the respiratory, reproductive, and skeletal organ systems, Pb disturbs the nervous system, while above a certain dose, both Cd and Pb become carcinogenic [3]. Other metals such as Cu, Cr, Ni, Zn have various biochemical and physiological functions, being essential elements. However, elevated concentrations of these elements become toxic [4][5][6]. High Cu concentration in the human body causes damage to internal organs and anemia, while high Cr concentration can affect the liver [7]. Moreover, the metals are non-biodegradable and persistent, and their high toxicity risk is related to the accumulation in soil, water, and living organisms [8]. As a consequence, maximum allowable limits are established for these elements in the environmental factors, and they should be removed from contaminated In all cases, the relative standard deviations (RSD%) for the parallel determinations were <5%. As presented in Table 1, the measured Si/Al ratio was >4 and the content of alkaline cations (Na + K) were higher than the content of Ca, indicating the presence of clinoptilolite as a major constituent in the analyzed volcanic tuff. The concentrations of major oxides were generally almost unaffected by crushing and heating treatment, and the differences between the contents of each oxide in different samples, expressed as Coefficients of Variation (CVs), were below 10%. A small decrease in Al 2 O 3 concentrations is observed in the samples treated at 550 • C, indicating the start of a dealumination process [26].
Dealumination (removing Al atoms) from the zeolitic framework arises at hightemperature treatments, probably caused by the breaking of Si-O(H)-Al bonds [27]. Even though the dealumination mechanism has not been entirely understood, it was reported that the treatment of zeolite at high temperatures leads to the formation of silanol defects. The term "defect" in a zeolite framework refers to the presence of silanol (SiOH) groups [28]. Under severe conditions, even Si atoms can be removed together with the Al atoms from the framework leading to structural defects that reduce the framework stability and thus in amorphization of the structure [28].
Molecules 2022, 27, x FOR PEER REVIEW 3 of 20 In all cases, the relative standard deviations (RSD%) for the parallel determinations were <5%. As presented in Table 1, the measured Si/Al ratio was >4 and the content of alkaline cations (Na + K) were higher than the content of Ca, indicating the presence of clinoptilolite as a major constituent in the analyzed volcanic tuff. The concentrations of major oxides were generally almost unaffected by crushing and heating treatment, and the differences between the contents of each oxide in different samples, expressed as Coefficients of Variation (CVs), were below 10%. A small decrease in Al2O3 concentrations is observed in the samples treated at 550 °C, indicating the start of a dealumination process [26].
Dealumination (removing Al atoms) from the zeolitic framework arises at high-temperature treatments, probably caused by the breaking of Si-O(H)-Al bonds [27]. Even though the dealumination mechanism has not been entirely understood, it was reported that the treatment of zeolite at high temperatures leads to the formation of silanol defects. The term "defect" in a zeolite framework refers to the presence of silanol (SiOH) groups [28]. Under severe conditions, even Si atoms can be removed together with the Al atoms from the framework leading to structural defects that reduce the framework stability and thus in amorphization of the structure [28].
According to XRD analysis, the used zeolitic tuff contains Ca-clinoptilolite (PDF 00-047-1870) as the main crystalline phase, accompanied by muscovite (PDF 00-007-0025), quartz (PDF 00-005-0490) and albite (PDF 00-019-1184) ( Figure 1). The low amorphous content attributed to the presence of quartz and kaolinized volcanic ash tuff is indicated by the hump in the region 2θ = 18-25° [28,29]. The applied thermal treatment led to a slightly decrease of the clinoptilolite diffraction lines intensities, and thus to a decreasing of crystallinity degrees of the zeolite samples, mainly for the samples with small particle size (0-0.05 mm), as follows: NZ1 (69.1%) > NZ1-200 (67.0%) > NZ1-350 (64.7%) > NZ1-550 (54.4%), and NZ2 (68.5%) > NZ2-200 (67.2%) > NZ2-350 (64.3%) > NZ2-550 (48.4%), respectively. Similar crystallinity loss by thermal treatment was reported for a natural clinoptilolite zeolite from Turkey [29]. The structure damage at 550 °C is indicated by the decrease of clinoptilolite peaks, which are more visible for the main peak at 22.4°. However, the complete amorphization of the crystalline structure does not occur until 550 °C. The thermal stability up to 550 °C of the investigated zeolitic tuff sample can be attributed to the reversible dehydration that arises with slight or no modification of the crystal structure [30]. At the same time, in all cases, the intensity and position of the peak attributed to crystalline quartz (SiO2) at 2θ = 26.6° did not change with increasing thermal treatment temperature. 10 15 20 25 30 35 40  The applied thermal treatment led to a slightly decrease of the clinoptilolite diffraction lines intensities, and thus to a decreasing of crystallinity degrees of the zeolite samples, mainly for the samples with small particle size (0-0.05 mm), as follows: NZ1 (69.1%) > NZ1-200 (67.0%) > NZ1-350 (64.7%) > NZ1-550 (54.4%), and NZ2 (68.5%) > NZ2-200 (67.2%) > NZ2-350 (64.3%) > NZ2-550 (48.4%), respectively. Similar crystallinity loss by thermal treatment was reported for a natural clinoptilolite zeolite from Turkey [29]. The structure damage at 550 • C is indicated by the decrease of clinoptilolite peaks, which are more visible for the main peak at 22.4 • . However, the complete amorphization of the crystalline structure does not occur until 550 • C. The thermal stability up to 550 • C of the investigated zeolitic tuff sample can be attributed to the reversible dehydration that arises with slight or no modification of the crystal structure [30]. At the same time, in all cases, the intensity and position of the peak attributed to crystalline quartz (SiO 2 ) at 2θ = 26.6 • did not change with increasing thermal treatment temperature.
The concentrations of heavy metals (Cu, Cd, Cr, Ni, and Pb) were measured in zeolite solid samples after acid microwave digestion, and those released in water at neutral pH (ratio zeolite: water = 1:10), measured by ICP-OES are presented in Table 2. The Cr concentrations in the volcanic tuff samples were below the limit of quantification (LOQ). The Cu concentrations were in the range of 1.65-2.03 mg kg −1 , Cr concentrations were in the range of 7.31-8.41 mg kg −1 , and the concentrations of Ni ranged between 3.21-4.36 mg kg −1 , while the concentrations of Pb ranged between 4.96-6.78 mg kg −1 .
In all cases, the CVs for concentrations of each individual metal in different samples were below 10%, while these variations do not show an increasing or decreasing tendency, indicating that they are caused rather by the measurement uncertainty of the analytical method and not due to particle sizes or temperatures applied during the thermal treatment. The concentrations of all the analyzed heavy metals leached from zeolites into ultrapure water at a neutral pH (pH ∼ = 7) were below LOQs (4 µg L −1 for Cu, 2 µg L −1 for Cd, 7 µg L −1 for Cr, 8 µg L −1 for Ni and Zn, and 10 µg L −1 for Pb). This is an important behavior when accounting for the heavy metals concentrations measured in the solution resulted from batch experiments and shows that the studied zeolites do not release these metals in contact with water, in the used experimental conditions. CEC values measured by the AMAS method varied in the range of 82.8-112.9 meq 100 g −1 and showed a small decrease with the increase in the thermal treatment temperature ( Table 3). The exchangeable Ca 2+ , the main component of the total CEC values determined by the AMAS method, is followed by exchangeable K + and lower quantities of Na + and Mg 2+ . However, considering the total amounts of Na + , K + , Ca 2+ and Mg 2+ measured after microwave acid digestion, it can be observed that Na + is the most mobile cation, with 84-99% of the total concentration in exchangeable form. The percent of exchangeable Ca 2+ is in the range of 47-83% of the total amount, the percent of exchangeable K + is in the range of 40-75% of the total amount, while only a very small part of the Mg 2+ is exchangeable (2.7-7.8%). As a general remark, the increasing temperature used for the zeolites treatment slowly decreased the exchangeability rate of these cations. The theoretical CEC values calculated based on the microwave-assisted acid extractions (242 meq 100 g −1 (NZ1), 231 meq 100 g −1 (NZ1-200), 246 meq 100 g −1 (NZ1-350), 227 meq 100 g −1 (NZ1-550) and, respectively, 220 meq 100 g −1 (NZ2), 221 meq 100 g −1 (NZ2-200), 230 meq 100 g −1 (NZ2-350), 248 meq 100 g −1 (NZ2-550)) highly exceeded the effective CEC value determined by AMAS method. These results indicate that 35-51% of exchangeable sites are active and can be implied in the exchange processes. According to the International Union of Pure and Applied Chemistry (IUPAC), the adsorbents are classified based on their sizes of pores into three categories: macropores, with a dimension greater than 50 nm, mesopores with dimensions ranging between 2 and 50 nm and micropores with dimension less than 2 nm [25].
As presented in Table 3, all the analyzed samples contain only mesopores. The surfaces of these pores are linked with active functional groups that contribute to the adsorption process and offer spaces for the sequestration of heavy metals. The surface areas ranged between 33-38 m 2 g −1 showing a slight diminishing in samples treated at higher temperatures. The specific surface area of porous materials is divided into external and internal specific surface areas [31]. By grinding, the external surface area increases, but the internal surface area, which has the most important contribution to the specific surface area in porous materials, is unchanged or even decreases. Thus, if the external surface area increases by grinding, it does not necessarily lead to an increase in the total specific surface area (determined by BET). Similarly, Burris and Juenger [32] reported that the specific surface area does not significantly increase by zeolite milling.

Zeolite Selectivity for Heavy Metals Sorption and Effect of Contact Time
The selectivity of used natural zeolite in our experiments for the sorption of studied heavy metals decreased as follows: Pb > Cr > Cu > Zn > Cd > Ni. Zamzow et al. [33] reported for the sodium form of clinoptilolite the following order of selectivity: Pb > Cd > Cs > Cu > Co > Cr > Zn > Ni > Hg, which, except Cd, is similar to our results. Belova [14] reported a sorption capacity for natural zeolite from the Yagodnisky deposit in the order Cu > Fe > Ni > Co. The natural zeolites studied by Sprynskyy et al. [34] adsorbed heavy metal ions in the following order: Pb > Cu > Cd > Ni, which is similar to our results. In addition, Hong et al. [35] reported an adsorption capacity trend: Pb > Cu > Ni. According to these findings, it can be concluded that the affinity of clinoptilolite natural zeolite toward different heavy metals is mostly similar, but for some cations, this is related to the zeolites' specific properties [36].
Generally, the metals' sorption reached equilibrium even after 5 min of contact time, mainly for the small particle size. Thus, a contact time of 60 min was considered long enough to compare the adsorption behavior of the two types of particle sizes (NZ1 and NZ2).

Influence of Zeolite Particle Size on Removal Efficiency (E%)
As shown in Figures 2-4, the heavy metals removal efficiencies are strongly influenced by the size of the adsorbent particle. Thus, when the NZ2 samples (particle size 0-0.05 mm) were contacted with the initial solution containing 5 mg L −1 heavy metals, the removal efficiencies (E%) reached nearby 90-99% after 5 min contact time, except for Ni 2+ for which E% was in the range of 42.4-72.4%. In the case of NZ1 with a particle size of 1-3 mm contacted with the same initial concentration (5 mg L −1 ), the removal efficiencies (E%) were generally much lower, only Pb being highly removed from solutions. olecules 2022, 27, x FOR PEER REVIEW 6 of 20 3 mm contacted with the same initial concentration (5 mg L −1 ), the removal efficiencies (E%) were generally much lower, only Pb being highly removed from solutions.
(a) (b)     According to the results, the removal strongly depends on the particle size of the zeolite, with fine particles much more efficient than coarse particles, because of a greater external surface area available for the sorption of heavy metals, which enhances the adsorption process. Even if the total specific surface area measured by BET does not increase notably for fine particles, the decrease in particle size leads to the increase of external surface area available for the interaction with metal ions in solution and results in shorter diffusion path lengths for sorption. The diffusion of metal ions from the surface to interparticle sites in zeolites is slow due to the interaction of metal ions with the surface functional group (electrostatic attraction, bond formation, etc.) [37].    According to the results, the removal strongly depends on the particle size of the zeolite, with fine particles much more efficient than coarse particles, because of a greater external surface area available for the sorption of heavy metals, which enhances the adsorption process. Even if the total specific surface area measured by BET does not increase notably for fine particles, the decrease in particle size leads to the increase of external surface area available for the interaction with metal ions in solution and results in shorter diffusion path lengths for sorption. The diffusion of metal ions from the surface to interparticle sites in zeolites is slow due to the interaction of metal ions with the surface functional group (electrostatic attraction, bond formation, etc.) [37].
Furthermore, the diffusion path length of the cations through the adsorbent cations is essentially shortened, which also simplifies the sorption [38]. Pernyeszi et al. [39] also reported that smaller adsorbent particle sizes have better sorption than larger particles due to a synergy between increased specific surface area and active sites.
The removal efficiency (E%) was remarkably increased when a smaller particle size (0-0.05 mm) zeolite was used. For the solution with an initial concentration of 5 mg L −1 , E% increased by 1.3, 5.7, 5.6, 3.6, 1.5, and 3.4 times for Pb, Cd, Zn, Cu, Cr and Ni. For the  Furthermore, the diffusion path length of the cations through the adsorbent cations is essentially shortened, which also simplifies the sorption [38]. Pernyeszi et al. [39] also reported that smaller adsorbent particle sizes have better sorption than larger particles due to a synergy between increased specific surface area and active sites.
The removal efficiency (E%) was remarkably increased when a smaller particle size (0-0.05 mm) zeolite was used. For the solution with an initial concentration of 5 mg L −1 , E% increased by 1.3, 5.7, 5.6, 3.6, 1.5, and 3.4 times for Pb, Cd, Zn, Cu, Cr and Ni. For the solution with an initial concentration of 10 mg L −1 , E% was improved on average by 1.9, 5.2, 5.7, 4.8, 2.5, and 3.0 times. When the initial concentration was 30 mg L −1 , E% was improved on average by 2.9, 6.1, 3.4, 2.9, 1.9, and 1.8 times.
The increase of E% due to the use of fine particle size is clearer at a shorter contact time (5 min). This implies that the fine particle sizes increase the sorption speed, and the system reaches an equilibrium state after a much shorter contact time compared with the larger particles.

Influence of Initial Heavy Metals Concentration in Solution
Concentration is one of the important factors that influence the removal of metal ions. As shown in Figure 2, the heavy metals removal efficiencies are influenced by the initial concentrations of heavy metals, as an effect of saturation of active sites from zeolite surfaces. However, even if the removal efficiencies generally decrease with the increasing initial concentrations, in terms of amounts of adsorbed metals, it can be observed that the increased metal concentration leads to a higher sorption capacity. This can be explained by the higher gradient of concentration between the solution and the adsorbent phase and by the higher number of cations around the active sites of the adsorbent, which increase the probability of adsorption [40]. Taamneh and Sharadqah [40] also reported that the percentage of adsorption of Cd and Cu on natural zeolite decreases by increasing the concentrations of the initial ions.

Influence of Thermal Treatment Temperature on Removal Efficiency (E%)
The influence of thermal treatment temperature on zeolite removal efficiency can be observed in Figures 2-4. In general, no relevant changes were observed in the removal efficiency for the treatment temperatures of 200 • C and 350 • C. At a higher temperature of 550 • C, small decreases in the removal efficiencies were observed mainly at shorter contact time. These results are correlated with the decreasing of clinoptilolite content observed in the XRD analysis of samples treated at 550 • C.

Amounts of Heavy Metals Ions Sorption from Contaminated Solutions
The evolution of the metal amounts in the adsorbent phase during the 60 min of experiments from the solutions with different initial concentrations are presented in Figures 5-7.

Amounts of Heavy Metals Ions Sorption from Contaminated Solutions
The evolution of the metal amounts in the adsorbent phase during the 60 min of experiments from the solutions with different initial concentrations are presented in Figures 5-7.    The amounts of heavy metals sorbed into the adsorbent phase, qe (mg g −1 ) calculated using Equation (1) are strongly influenced by the concentrations of heavy metals in the initial solutions (Figures 5-7). In addition, the zeolites particle sizes and metals species play an important role in qe values. When a solution containing 5 mg L −1 of each heavy metal made contact with the particle size 0-0.05 mm zeolites, the qe values ranged between 0.0488-0.0500 mg g −1 Pb, 0.0316-0.0498 mg g −1 Cd, 0.0370-0.0500 mg g −1 Zn, 0.0490-0.0500 mg g −1 Cu, 0.0479-0.0500 mg g −1 Cr, and 0.0212-0.0495 mg g −1 Ni. Almost the entire amounts of all the heavy metals from solution were sorbed onto the zeolites after the 60 min contact time. The solution containing 5 mg L −1 heavy metals mixed with the zeolite with a 1-3 mm particle size led to the qe values on a more dispersed domain: 0.0168-0.0492 mg g −1 Pb, 0.0023-0.0244 mg g −1 Cd, 0.0022-0.0310 mg g −1 Zn, 0.0038-0.0383 mg g −1 Cu, 0.0151-0.0476 mg g −1 Cr, and 0.0037-0.0244 mg g −1 Ni. In this case, the qe values were significantly influenced by the contact time.
In the case of solution with an initial concentration of 10 mg L −1 heavy metals in contact with a 0-0.05 mm particle-size zeolites, the qe values ranged between 0.0978-0.0998 mg g −1 Pb, 0.0424-0.0969 mg g −1 Cd, 0.0482-0.0926 mg g −1 for Zn, 0.0946-0.0999 mg g −1 Cu, 0.0944-0.0999 mg g −1 for Cr, and 0.0307-0.0678 mg g −1 for Ni. Almost the entire amounts of heavy metals from the solution are adsorbed onto the zeolites after the 60 min contact time. The values of qe are, in general, twice that of those in the experiment with the 5 mg L −1 initial concentration. When the 10 mg L −1 solution was mixed with the 1-3 mm particle-size zeolites resulted in qe values in the ranges of 0.0142-0.0978 mg g −1 Pb, 0.0052-0.0264 mg g −1 Cd, 0.0035-0.0326 mg g −1 Zn, 0.0068-0.0460 mg g −1 Cu, 0.0205-0.0707 mg g −1 Cr, and 0.0084- The amounts of heavy metals sorbed into the adsorbent phase, q e (mg g −1 ) calculated using Equation (1) are strongly influenced by the concentrations of heavy metals in the initial solutions ( Figures 5-7). In addition, the zeolites particle sizes and metals species play an important role in q e values. When a solution containing 5 mg L −1 of each heavy metal made contact with the particle size 0-0.05 mm zeolites, the q e values ranged between 0.0488-0.0500 mg g −1 Pb, 0.0316-0.0498 mg g −1 Cd, 0.0370-0.0500 mg g −1 Zn, 0.0490-0.0500 mg g −1 Cu, 0.0479-0.0500 mg g −1 Cr, and 0.0212-0.0495 mg g −1 Ni. Almost the entire amounts of all the heavy metals from solution were sorbed onto the zeolites after the 60 min contact time. The solution containing 5 mg L −1 heavy metals mixed with the zeolite with a 1-3 mm particle size led to the q e values on a more dispersed domain: 0.0168-0.0492 mg g −1 Pb, 0.0023-0.0244 mg g −1 Cd, 0.0022-0.0310 mg g −1 Zn, 0.0038-0.0383 mg g −1 Cu, 0.0151-0.0476 mg g −1 Cr, and 0.0037-0.0244 mg g −1 Ni. In this case, the q e values were significantly influenced by the contact time.
In the case of solution with an initial concentration of 10 mg L −1 heavy metals in contact with a 0-0.05 mm particle-size zeolites, the q e values ranged between 0.0978-0.0998 mg g −1 Pb, 0.0424-0.0969 mg g −1 Cd, 0.0482-0.0926 mg g −1 for Zn, 0.0946-0.0999 mg g −1 Cu, 0.0944-0.0999 mg g −1 for Cr, and 0.0307-0.0678 mg g −1 for Ni. Almost the entire amounts of heavy metals from the solution are adsorbed onto the zeolites after the 60 min contact time. The values of q e are, in general, twice that of those in the experiment with the 5 mg L −1 initial concentration. When the 10 mg L −1 solution was mixed with the 1-3 mm particle-size zeolites resulted in q e values in the ranges of 0.0142-0.0978 mg g −1 Pb, 0.0052-0.0264 mg g −1 Cd, 0.0035-0.0326 mg g −1 Zn, 0.0068-0.0460 mg g −1 Cu, 0.0205-0.0707 mg g −1 Cr, and 0.0084-0.0309 mg g −1 Ni, with increasing q e values in time.

Isotherm and Kinetic Modeling
The sorption data of Pb, Cd, Zn, Cu, Cr and Ni ions onto zeolite with a 0-0.05 mm particle size thermally treated at 200 • C were analyzed using the nonlinear forms of Langmuir and Freundlich isotherm models. The nonlinear plots of Langmuir and Freundlich isotherm models for Pb, Cd, Zn, Cu, Cr and Ni ions sorption onto zeolite with a 0-0.05 mm particle size thermally treated at 200 • C are presented in Figure 8. The Langmuir and Freundlich isotherm parameters are given in Table 4.  The Langmuir and Freundlich isotherm parameters are given in Table 4. The best fit was achieved with the Langmuir equation according to the high values of correlation coefficients (R 2 ) obtained for Cd (R 2 = 0.9615), Zn (R 2 = 0.9999), Cu (R 2 = 0.9814), Cr (R 2 = 0.9812) and Ni (R 2 = 0.9930) sorption. In the case of Pb sorption, the correlation coefficient of Langmuir isotherm is high (R 2 = 0.9982) and close to that of Freundlich isotherm (R 2 = 0.9989). The maximum adsorption capacities (q max) from Langmuir isotherm were 0.393 (Pb), 0.112 (Cd), 0.094 (Zn), 0.129 (Cu), 0.142 (Cr) and 0.069 mg g −1 (Ni), respectively. The following series was depicted based R 2 values obtained from the Langmuir model: Zn, Pb, Ni, Cu, Cr and Cd. The n values (Freundlich isotherm model) indicated normal sorption of the metals onto zeolite. According to the obtained R 2 values, the Freundlich isotherm model cannot adequately describe the sorption of Cd, Zn, Cu, Cr and Ni onto zeolite with a 0-0.05 mm particle size.
In addition, the experimental results were analyzed using the nonlinear forms of pseudo-first-order (PFO) and pseudo-second-order (PSO) kinetic models. The nonlinear plots of PFO and PSO for Pb, Cd, Zn, Cu, Cr and Ni ions sorption onto zeolite with a 0-0.05 mm particle size thermally treated at 200 • C are presented in Figures 9 and 10, respectively. The PFO and PSO parameters are given in Table 5 for all the studied metals. The experimental q e values (q e , exp ) were very close to the calculated q e, values (q e,calc ) for PFO and PSO. The R 2 values obtained from the PFO model were very high compared with the R 2 values obtained from the PSO model. Thus, the PFO model described the sorption data of Pb, Cd, Zn, Cu, Cr and Ni ions sorption onto the zeolite with a 0-0.05 mm particle size thermally treated at 200 • C better than PSO.
Heavy metal ions can be immobilized by zeolites by two main mechanisms: ionexchange and chemisorption [41]. Ion exchange involves the substitution of the exchangeable cations (Na + , K + , Ca 2+ and Mg 2+ ) located in the zeolite crystalline lattice by heavy metals cations from the solution. Since metals retained in zeolite by ion exchange are weakly physically bounded, they can be replaced to form zeolite with NH 4 + . Chemisorption is based on the formation of stable inner-sphere complexes by chemical bonds of functional groups (mainly OH − ) with metal ions outside the hydration envelope [41,42]. In zeolites, ion-exchange processes generally dominate chemisorption, but this depends on the metal species [42]. To evaluate the sorption of metals on zeolites due to the ion-exchange process, the zeolite used for metal ions removal in sorption experiments was mixed with ammonium acetate solution 1 M (ratio 1:50), and the desorbed metals in the extraction solutions were measured using ICP-OES. It was assumed that NH 4 + assures the removal of all metal ions retained by ion exchange, while the heavy metals that remained in the zeolites were retained by chemisorption [42]. According to the results presented in Table 6, the total immobilization of heavy metals in zeolite is caused both by ion exchange and chemisorption, but their contribution depends on the type of metal.
tively. The PFO and PSO parameters are given in Table 5 for all the studied metals. The experimental qe values (qe,exp) were very close to the calculated qe, values (qe,calc) for PFO and PSO. The R 2 values obtained from the PFO model were very high compared with the R 2 values obtained from the PSO model. Thus, the PFO model described the sorption data of Pb, Cd, Zn, Cu, Cr and Ni ions sorption onto the zeolite with a 0-0.05 mm particle size thermally treated at 200 °C better than PSO.   In the case of Cu, ion exchange is by far the most important mechanism of retention (over 85%). In the case of Pb and Cd, about 60-70% of retention is caused by ion exchange; in the case of Zn, the ion exchange and chemisorption have almost equal contributions to its retention, while in the case of Cr and Ni, the chemisorption retention mechanism prevails, but the ion exchange still contributes approximately 41-47% for Cr, and 33-43% for Ni. Krol et al. [41] reported that chemisorption predominates for Cr, chemisorption and ion exchange are almost of equal importance for Cd and Pb, whereas in the case of Ni ion-exchange processes prevail.      Table 6. Desorption rate (%) of Pb, Cd, Zn, Cu, Cr, and Ni from zeolite with a 0-0.05 mm and 1.0-3.0 mm particle size thermally treated at 200 • C, after the sorption experiments from solutions of 5 mg L −1 , 10 mg L −1 , 30 mg L −1 and 60 min contact time.

Materials
All chemicals used in this study were of analytical reagent grade. Emsure ® ACS premium-grade acids HNO 3 65%, HCl 37%, and HF 40%, purchased from Merck (Darmstadt, Germany), were used for the sample digestion. Ultrapure water (18 MΩ cm −1 ) obtained from a Millipore Direct Q3 (Millipore, Bedford, MA, USA) was used for dilutions. Standard solutions for external calibration of ICP-OES were prepared by the stepwise dilution of a Merck Millipore CertiPur ICP multi-elemental standard solution IV (23 elements) 1000 mg L −1 , purchased from Merck (Darmstadt, Germany). The accuracy of the analysis for total metal concentrations in zeolite samples was assessed using CRM BCS-CRM 375/1 soda feldspar (Bureau of Analyzed Samples, Middlesbrough, UK). The recovery percentages of the analyzed elements in CRM were in the range of 86-102%.
To determine the major elements (Al, Ca, Mg, K, Na, Fe, Mn) and trace metals (Cu, Cd, Cr, Ni, and Pb) in zeolite samples, microwave-assisted acid digestion with an Xpert system (Berghof, Eningen, Germany) was carried out. An amount of 1 g of zeolite sample was digested with a 14 mL mixture of HNO 3 65%:HCl 37%:HF 40% (3:9:2, v:v:v) in a three-steps heating program until 200 • C, with a total digestion time of 40 min. After cooling down at room temperature, 20 mL of saturated H 3 BO 3 were added, then heated again at 160 • C for 15 min, then filtered and diluted with ultrapure water to a final volume of 100 mL. Three parallel determinations were carried out for each sample analysis.
In order to evaluate the possible release of Cu, Cd, Cr, Ni, Zn, and Pb into the water at a neutral pH, 5 g of each zeolite sample was mixed with 45 mL ultrapure water and stirred at room temperature (23 ± 2 • C) for 60 min, filtered through a cellulose filter, then analyzed for metals. The resulting solutions (from zeolite digestion, released into ultrapure water and from batch adsorption experiments) were analyzed for metals using a dual-viewing inductively coupled plasma optical emission spectrometer Optima 5300DV (Perkin Elmer, Norwalk, CT, USA). The seven-point linear calibration curves over the range 0-10 mg L −1 element were constructed for each analyzed element. The concentrations of major elements (Al, Ca, Mg, K, Na, Fe, Mn) in the zeolites were converted to oxides using atomic and molecular masses. The SiO 2 content was determined gravimetrically [26].
The cation exchange capacity (CEC) was determined by measuring using ICP-OES the major cations (K, Na, Ca, and Mg) extracted in ammonium acetate solution 1 M (AMAS method). Total surface area, pore radius and total pore volumes were obtained from N 2 adsorption-desorption isotherms using the Brunauer-Emmett-Teller (BET) method for total surface area evaluation and Dollimore-Heal model for porosity data. The isotherms were obtained using a Sorptomatic 1990 apparatus (Thermo Electron Corporation, Waltham, MA, USA). The X-ray diffraction (XRD) patterns were recorded at room temperature using a D8 Advance (Bruker, Karlsruhe, Germany) diffractometer with CuKα radiation (λ = 1.54060 Å), operating at 40 kV and 40 mA.

Batch Sorption Experiments
A multicomponent stock solution containing Cu, Cd, Cr, Ni, Zn and Pb 1 g L −1 was used to prepare the chosen concentrations by dilution with ultrapure water. The experiments were performed in batch mode, contacting 5 g of zeolites (NZ1-200, NZ1-350, NZ1-550, and respectively, NZ2-200, NZ2-350, NZ2-550) with 45 mL solutions at different Cu 2+ , Cd 2+ , Cr 3+ , Ni 2+ , and Pb 2+ initial concentrations (5 mg L −1 , 10 mg L −1 ; 30 mg L −1 ) at a stirring rate of 100 rpm. All the experiments were performed at room temperature (23 ± 2 • C). Samples were taken at established time intervals (5, 10, 30, 60 min) and were filtered before ICP-OES analysis. All the experiments were carried out in triplicate, and the average values were used to calculate the results. The standard deviations were determined at less than 5%.
The heavy metal amounts in the adsorbent phase, q e (mg g −1 ) were calculated using Equation (1), while the removal efficiency, E (%), was calculated using Equation (2) [43]: (1) where q e is the heavy metal amounts adsorbed per gram of adsorbent at equilibrium (mg g −1 ), V is the volume of solution (mL), m is the weight of zeolite (g), C e is the equilibrium metals concentrations (mg L −1 ) and C o is the initial metal concentrations (mg L −1 ). The heavy metal amounts desorbed from zeolite resulted from sorption experiments were measured after desorbing in ammonium acetate solution 1 M and the zeolite: solution ratio of 1:50 [42]. The resulted slurries were mixed at a stirring rate of 100 rpm for 2 h at room temperature, then were centrifuged and filtered. The metal concentrations in the resulted solutions were determined using ICP-OES. Finally, the quantities of metals desorbed from the solid phase were calculated considering the initial amount of zeolite (2 g) and the final volume of the extraction solutions (100 mL).

Adsorption Isotherms and Kinetics
The Langmuir and Freundlich isotherm models were applied to describe the sorption process of Pb, Cd, Zn, Cu, Cr and Ni ions onto the zeolite with a 0-0.5 mm particle size thermally treated at 200 • C and the pseudo-first-order and pseudo-second-order kinetic models were applied to study the adsorption kinetics. The nonlinear forms of the considered models are given in Table 7 [44]. The Langmuir isotherm suggests monolayer sorption onto an adsorbent surface [44]. The Freundlich isotherm suggests that the adsorption occurs onto a heterogeneous surface [44]. The n value indicates a normal adsorption when n < 1, cooperative adsorption when n > 1 and a favorable adsorption when 1< n < 10 [45]. PFO, proposed by Lagergren, assumes physical adsorption as the rate-controlling mechanism, while PSO assumes that the chemisorption controls the reaction rate [46][47][48]. Table 7. Nonlinear equations of Langmuir and Freundlich isotherms, PFO and PSO kinetic models.

Model Equations Description Reference
Langmuir isotherm q e = q max K L C e 1+K L C e q max is the maximum adsorption capacity (mg g −1 ) and K L is the Langmuir constant (L mg −1 ) [44,48] Freundlich isotherm q e = K F C e (1/n) K F is related to adsorption capacity (mg 1−1/n L 1/n g −1 ) and 1/n is the adsorption intensity [44,48] PFO q t = q e 1 − e −k 1 t q t is the amount adsorbed at time t (mg g −1 ) and k 1 is the first-order rate constant (min −1 ) [44,48] PSO q t = q 2 e k 2 t 1+q e k 2 t k 2 is the second-order rate constant (g mg·min −1 ) [44,48] The nonlinear regression was performed using OriginPro software, version 2020b, OriginLab Corporation, Northampton, MA, USA.

Conclusions
The simultaneous removal of heavy metals (Cu, Cd, Cr, Ni, Zn and Pb) from contaminated aqueous solutions using thermally treated zeolitic volcanic tuffs from Racos , (Romania) as adsorbent was studied. The selectivity of the natural zeolite (Ca-clinoptilolite type) for the sorption of the studied heavy metals cations followed, in general, the following trend: Pb > Cr > Cu > Zn > Cd > Ni. Heavy metals removal efficiencies were powerfully influenced by the particle size, the samples with smaller particle sizes (0-0.05 mm) being more efficient in metal removal. The amount of metal adsorbed per unit of zeolite mass increased with the metal concentration in the initial solution. Generally, no relevant changes were observed in the removal efficiency of the zeolitic volcanic tuffs thermally treated at 200 • C and 350 • C. Though, in the case of samples thermally treated at 550 • C, a small decrease in the removal efficiencies was observed. The Cd, Zn, Cu, Cr, Zn, and Ni ions sorption was best described by the Langmuir model according to the high values of the correlation coefficient. The results indicated that the pseudo-first-order kinetic model presented the best correlation to the experimental data. The total immobilization of heavy metals in zeolite is caused mainly by ion exchange, but also by chemisorption. The obtained results demonstrate that the zeolitic volcanic tuff from Racos , (Romania) is an efficient adsorbent for the removal of heavy metals from aqueous solutions, with removal efficiencies reaching 99%, in specific conditions of initial concentration in the solution, particle sizes and contact time.