Equilibrium ion exchange studies of Ni2+ on homoionic forms of clinoptilolite

A natural zeolite (clinoptilolite) that is mined in KwaZulu-Natal, South Africa, was evaluated for the removal of Ni2+ from wastewater. In particular, the effect of zeolite modification on Ni2+ removal from synthetic wastewater was investigated. The natural clinoptilolite was pretreated with 2 M metal chlorides for 24 h to yield near homoionic Na+, K+ and Ca2+ forms. A comparison of the isotherms for the Na+-Ni2+, K+-Ni2+, Ca2+-Ni2+ and natural-Ni2+ systems gave insight into how the displaced ion affects the selectivity of the clinoptilolite for Ni2+. The Na+, K+ and natural forms show highly selective convex isotherms whereas the Ca2+ form has a concave graph suggesting that the selectivity series is Ca2+> Ni2+> (Na+, K+, natural). Thermodynamic properties revealed that the Ni2+ sorption capacity increases as the values of the equilibrium constant and Gibbs free energy increase with increasing temperature from 298.15 K to 348.15 K. The enthalpy change was positive for all forms of clinoptilolite; values of 26.00 kJ/mol, 18.72 kJ/mol and 42.05 kJ/mol were obtained for exchange of Ni2+ into Na+, K+ and Ca2+ forms, respectively. The positive changes in enthalpy provide an indication that the sorption reaction is endothermic for Ni(II). The Gibbs free energy values were all negative except for Ca2+-exchanged clinoptilolite at 298.15 K and 308.15 K, for which the values were positive 3.10 kJ/mol and 0.53 kJ/mol, respectively. The entropy values for Ni2+ sorption were also positive; values of 0.12 kJ/mol.K, 0.08 kJ/mol.K and 0.14 kJ/mol.K were obtained for the Na+, K+ and Ca2+ forms, respectively. As expected, the enthalpy obtained from the Van’t Hoff plot is dependent not only on the metal ion being adsorbed, but also on the ion being displaced. Pretreatment of the zeolite enhances the removal efficiency provided that monovalent ions are used for the pretreatment. Thus clinoptilolite is an effective low-cost absorbent for the removal of Ni2+ from aqueous solutions.


Introduction
Heavy metal pollution is an environmental problem of concern worldwide.The increasing levels of heavy metals in the environment represent a serious threat to human health and ecological systems.Soluble and mobile heavy metal species are non-biodegradable and tend to bioaccumulate in living organisms causing various diseases and disorders.][10] One of the important properties of zeolites is that they show selectivity in adsorption, i.e. they possess different affinities for different ions. 11,124][15][16][17][18][19] These authors pointed out that zeolite in homoionic forms exhibits a significantly increased ability to remove heavy metals from wastewater.NaCl is most often used as the pretreatment agent.Prior to any ion-exchange application, most exchangeable ions from the structure of the material are removed by pretreatment and replaced by more easily removable ones.Pretreatment of natural zeolites with, for example, acids, bases and surfactants, is also used to improve their ion-exchange capacity.Most pretreatment operations increase the content of a single cation, called a homoionic form.
The evaluation of the ion-exchange properties of zeolites is based on equilibrium data for a particular exchange reaction.On the basis of these data, the main thermodynamic properties, such as the equilibrium constant (K eq ) and Gibbs free energy ΔG˚, can be computed using a suitable model.The use of a reliable model for the exchange process is particularly important when one needs to predict the ion-exchange behaviour of the zeolites for varying compositions of the aqueous phase based on experimental data.[22][23][24] Fitting of adsorption isotherm equations to experimental data is often an important aspect of data analysis.[27] The Langmuir equation assumes that the adsorbed species forms a monolayer.But monolayer formation is possible only for a dilute solution.Under high concentration conditions the assumption is no longer valid as adsorbates accumulate to form multiple layers.The Langmuir equation also assumes that adsorbed molecules do not interact with each other laterally.This is impossible as weak forces of attraction exist even between molecules of the same type.Another assumption is that all the sites on the solid surface are equivalent in size and shape and have equivalent affinity for adsorbate molecules, i.e. the surface of the solid is homogeneous.But real solid surfaces are heterogeneous.Because clinoptilolite minerals have high surface irregularities, the adsorption models (Langmuir and Freundlich) should not be used to explain the adsorption equilibrium phenomenon. 28eolites have been found to be highly efficient in removing heavy metals from wastewater and the costs involved are still significantly below competing technologies.Although major breakthroughs have been made towards the use of zeolites in environmental remediation, most researchers have focused on heavy metal ions such as Pb 2+ , Cu 2+ and Co 2+ rather than on Ni 2+ removal.Two interesting observations with respect to Ni 2+ adsorption on clinoptilolite have been made by Sprynskyy et al. 29 After its initial fast uptake from mixed metal feed concentrations, the second phase of adsorption is characterised by desorption and is referred to as an 'inversion phenomenon'.It is thought that this phenomenon is caused by the counter-diffusion of displaced extra-lattice cations from the deeper layers of the zeolite.As high concentrations of these counter-ions diffuse out of the zeolite pores, the Ni 2+ is displaced and readsorbed.This displacement indicates that Ni 2+ adsorbs in the mesopores, rather than in the zeolite channels.It was observed by Sprynskyy et al. 29 that when metal ions like Pb 2+ are adsorbed on clinoptilolite, there is usually only a slight difference between the loading capacities of these metal ions when single component feeds are compared to mixed feeds.This finding implies that metal ions are sorbed on specific sites.However, with Ni 2+ , adsorption significantly decreases in mixed feeds as a consequence of competition with other metals.Sprynskyy et al. 29 concluded that the adsorption of Ni 2+ ions is not site specific.Moreover, the zeolite selectivity for Ni 2+ is generally low, hence improvements in Ni 2+ selectivity are important for industrial applications, especially when implementation of more stringent standards for discharge of heavy metals into receiving environments are taken into consideration.It has been proven that converting zeolite to homoionic forms improves selectivity, but we are not aware of any study which compares different homoionic forms for the removal of Ni 2+ from aqueous solutions.In this study, physical (i.e.pore diameter and volume) and thermodynamic differences between these forms have been found; such data have not been published before.

Theory: Ion-exchange isotherms
Ion-exchange isotherms are plots of the equilibrium fraction of an exchanging ion in solution against the equilibrium fraction of the same ion in the zeolite at the same temperature.The isotherms are plotted in terms of equivalent cation fraction of the ion in the solution (X sol ) against that in the solid (X zeo ) in accordance with the analysis of Inglezakis et al. 23 The equivalent cation fraction in solution and on the clinoptilolite is calculated by using Equations 3 and 4 given below.

Construction of ion-exchange isotherms
In general, the ion-exchange reaction between a solution containing the cation A zA+ (where A is a cation of valence z A ) and the B form of clinoptilolite (B being a cation of valence z B ) may be written as 30 : in which L is a portion of the clinoptilolite framework holding a unit negative charge and the subscript aq denotes the solution phase.The equivalent fraction of the exchanging cation in the solution (X sol ) is therefore given by: where m s A and m s B are the molarities of the ions A and B in solution, respectively.The equivalent cation fraction in the clinoptilolite is given by: where W is the zeolite mass in grams, V is the solution volume used in litres, M A,i and M A,f are the initial and final concentrations of the exchanging ion (in moles per litre) and CEC is the cation-exchange capacity of the zeolite (in eq/g). 23

Experimental methods
The raw zeolite sample used in this study was obtained from Pratley (Pty) Limited (Kenmare, South Africa) which mines the zeolite in KwaZulu-Natal (South Africa).All chemicals and reagents used for experiments and analysis were analytical grade supplied by Merck Ltd.

Zeolite preparation
Small grains of zeolite were obtained by first using a hammer to break the as-received samples into small pieces.These small pieces were then ground using a pestle and mortar and sieved to yield fractions differing in diameter from 0.60 mm to 0.85 mm.Sieving was repeated several times to minimise the retention of smaller grains in a sample with a larger size range.Prior to the batch adsorption experiments, the crushed zeolite was washed with distilled water three times to remove the surface dust, and then dried in an oven at 343.15 K for 24 h until a constant weight was attained.

Preparation of homoionic forms
Near homoionic forms of Na + , K + and Ca 2+ clinoptilolite were generated by treating 30-g batches of the purified clinoptilolite with 300 mL of 2 M chloride salts (NaCl, KCl and CaCl 2 for Na + , K + and Ca 2+ , respectively).The mixtures were then placed in a Labex shaking incubator (Edenvale, South Africa) at 298.15 K for 24 h at a speed of 200 rpm.The solutions of the Cl -salts were replaced with fresh ones for a further 24 h.The treated clinoptilolite grains were washed several times with distilled water to eliminate excess metal chlorides, and dried in an oven at 343.15 K for 24 h.Treated zeolite fractions were used as adsorbents for Ni 2+ removal.Metal chloride treatment was conducted based on the findings of previous studies that alkali and alkaline earth metals are cheap, commonly available and are the most effective exchangeable ions for heavy metal removal. 2

Equilibrium studies
Equilibrium studies were done as follows.A stock solution (1000 mg/L) of Ni 2+ was prepared by dissolving 2.04 g of NiCl 2 •6H 2 O in 1 L of distilled water.The composition of the synthetic aqueous solution used in this study was based on those used previously and the concentrations were within the range of typical industrial wastewaters, that is, 0.1 mg/L to 100 mg/L. 28nthetic samples were prepared to give Ni 2+ concentrations of 20 mg/L, 40 mg/L, 60 mg/L, 80 mg/L and 100 mg/L by adding an appropriate amount of NiCl 2 •6H 2 O stock solution to deionised water.The masses of clinoptilolite used in the experiments ranged from 0.2 g to 1.5 g and the solution volume ranged from 10 mL to 100 mL.Before adding the adsorbents, the pH of the solution was adjusted to 7, i.e. the pH at which H + and Ni 2+ competition is minimal, using either 0.1 M NaOH or 0.1 M HNO 3 solution, following the method of Gaus and Lutze 26 .
The zeolite mass to solution volume ratios and the aqueous mixture compositions used in the experiments were designed to yield a relatively evenly spaced distribution of points along the ion-exchange isotherm as well as significant differences in the initial and final concentrations of the cation in solution.No background electrolyte was added during the ion-exchange experiments.The metal concentration in the liquid phase was determined at the beginning (C o ) and at the end (C f ) of the adsorption.The following equation was used to compute the percentage uptake of the metal by the clinoptilolite:

Results
The chemical analysis of the natural and homoionic clinoptilolite samples is given in Table 1.In the natural sample about 50% of the exchangeable ions were K + .Complete exchange of cations was impossible to achieve, especially for the Na + -and Ca 2+ -treated clinoptilolite as it is assumed that K + present in the clinoptilolite did not exchange significantly with other cations.This behaviour is attributed to the location of K + .Based on the chemical composition of the natural clinoptilolite used in this study, the molecular formula for the natural zeolite was: (Na The pretreatment process led to the production of different forms of zeolite depending on the treatment agent used.The molecular formulae of the pretreated samples were as follows: Na + -clinoptilolite: (Na The cation-exchange capacity (CEC, milliequivalents (meq) per gram) of the zeolite samples was derived from the analysis given in Table 1.
The calculation of CEC was based on the assumptions that (1) Al 3+ and Fe 3+ substitute for Si 4+ in the tetrahedral sites and result in a negatively charged structure, (2) this negative charge is balanced by the alkali and alkaline earth ions in the intracrystalline cation-exchange sites and (3) any other exchangeable ions present in a homoionic form of clinoptilolite are assumed to occupy inaccessible sites; for instance, the K + , Ca 2+ and Mg 2+ present in the Na + -clinoptilolite are in inaccessible exchange sites and do not participate in the ion-exchange process.An additional assumption is that negligible mineral impurities are present in the sample. 22perimental CECs of clinoptilolite were determined and were found to range from 1.38 meq/g to 2.29 meq/g for all forms of clinoptilolite (for natural and pretreated clinoptilolite).
Previous studies showed that the South African zeolite comprises mainly clinoptilolite (80-85%) and coexists with impurities of opaline, cristobalite, K-feldspar and traces of sanidine. 2From the X-ray flourescence results, the Si:Al ratio was calculated to be 5.14.The chemical composition, the theoretical exchange capacity and the Si:Al ratio (generally ranging from 4 to 5.5) are typical for clinoptilolite. 31Low SiO 2 members are enriched with Ca 2+ , whereas high SiO 2 clinoptilolite is enriched with K + , Na + and Mg 2+ .It was found that the natural zeolite was predominantly clinoptilolite.

Effects of ion exchange
The surface area and pore volume data for Na + , K + and Ca 2+ form clinoptilolite are presented in Table 2.The ionic radii taken from Cotton and Wilkinson 32 have also been included in Table 2.The data for the Na + and Ca 2+ forms are within experimental error, while the K + form has a higher surface area and pore volume.

Note:
The surface area and pore volume data were obtained from Brunauer Emmett Teller (BET) analysis.The ionic radii are values from Cotton and Wilkinson 32 .
Figures 1 to 4 present the experimental ion-exchange results.The initial and final concentration of Ni 2+ was measured as well as the concentration of the displaced ions, Na + for the Na + form, K + for the K + form and Ca 2+ for the Ca 2+ form.These measured concentrations, zeolite masses and the volumes of solutions were used in the construction of ion-exchange isotherms.do not proceed to completion (that is, do not attain X solution = 1 for X zeolite = 1).This result is attributed to the fact that in these systems only a fraction of the total CEC is available to the incoming cations as a result of crystallite occlusion; nonetheless the CEC is close to 1 (0.9±0.1).The same observation was reported by Breck 33 .When the isotherm exhibits a convex profile, e.g. for Na + -Ni 2+ , K + -Ni 2+ and natural-Ni 2+ systems, the uptake of Ni 2+ from the solution is known to follow a Langmuir-type isotherm. 34The Na + , K + and natural forms show highly selective convex graphs, whereas the Ca 2+ form has a concave graph, suggesting the selectivity series to be Ca 2+ > Ni 2+ > (Na + , K + , natural).This selectivity series is in agreement with the investigation done by Colella 34 on Italian clinoptilolite.Figure 4 demonstrates that the clinoptilolite has a greater affinity for Ca 2+ than for Ni 2+ .This observation is consistent with the expectation that the Ca 2+ ion has a relatively lower desorption ratio than Na + and K + because of the strong interaction (ionic in nature) between clinoptilolite and Ca 2+ (it is difficult to destroy the Ca 2+ -O-Al bonds).A similar observation was made by Ćurković et al. 35 for Serbian clinoptilolite.A remarkably high selectivity of natural-, K + -and Na + -exchanged clinoptilolite for Ni 2+ is observed.Percentage error bars for some points are indicated.The curve was fitted to the isotherm data using a polynomial of order 2.
In order to obtain the parameter capable of describing the selectivity/ non-selectivity of a given form of zeolite, the thermodynamic equilibrium constant (K eq ) was evaluated using the following procedure outlined by Khan and Singh 36 : where a Z denotes the activity of adsorbed Ni 2+ , a S is the activity of Ni 2+ in solution at equilibrium, C z is the milligrams of Ni 2+ adsorbed per litre of solution in contact with the clinoptilolite surface, C s denotes the milligrams per litre of Ni 2+ in solution at equilibrium, f z is the activity coefficient of the adsorbed Ni 2+ and f s is the activity coefficient of the which reduces to Equation 9 The values of K eq were calculated as the ratio of the equilibrium concentration of Ni 2+ on the zeolite and in solution attained after 24 h of adsorption.The standard free energy change on adsorption (ΔG°) was calculated using the following equations: ΔG° = -RT lnK eq Equation 11 Several researchers have reported ΔG° values for adsorption in zeolites. 23,25In their calculations they only reported ΔG°a ds , and did not mention the ΔG°d es of the ion being displaced.By considering these two parameters, a true Gibbs free energy of exchange (ΔG°e xc ) can be computed which is representative of the whole system.Therefore Equation 12 is used to calculate the standard free energy of reaction: ΔG°e xc = ΔG°a ds -ΔG°d es Equation 12where ΔG°e xc denotes the Gibbs free energy of ion-exchange reaction, ΔG°a ds is the Gibbs free energy of adsorption of Ni 2+ onto clinoptilolite and ΔG°d es is the Gibbs free energy of desorption of exchangeable ions in the clinoptilolite (Na + , K + and Ca 2+ ).R is the universal gas constant (8.314J/mol.K) and T is the temperature in Kelvin.
The ΔS° and ΔH° values were obtained from the slope and intercept of the Van't Hoff plots (plots of the natural logarithm of equilibrium constant of the reaction versus the reciprocal of temperature in Kelvin).
lnK eq = ΔS°/R -(ΔH°/R)1/T Equation 13The enthalpy of the reaction can be found from the gradient of the plot, which equals -ΔH°/R, and the entropy is the intercept, ΔS°/R.Table 3 reports the thermodynamic values at different temperatures.
The equilibrium constants K eq derived from this study were quite high compared to those reported by Argun 27 despite using the same type of zeolite.Argun reported very low K eq values of 3.28, 2.97 and 2.65 at 293.15 K, 313.15K and 333.15 K, respectively, using Turkish natural clinoptilolite from the Langmuir isotherm constants to approximate the equilibrium constant, whereas, in this study, we found K eq to be 23.69 at 298.15 K, which increased as temperature was increased (Table 3).A re-evaluation of the isotherm data in Argun's 27 study (after 3-h contact time instead of 60 min as in the original paper) at 293.15 K and at an initial concentration of Ni 2+ of 25 mg/L for 1 g of clinoptilolite yielded a K eq of 18 for the natural-Ni 2+ system, which is very close to the values obtained in this study.The recalculated ΔG° value at 293.15 K is -7.04 kJ/mol (from -2.89 kJ/mol).Our results demonstrate that the reliance on linearised Langmuir equations potentially limits the ability to model sorption data accurately.
The thermodynamic quantities reported in Table 3 are in accordance with the selectivity series depicted from the isotherm plots.The Gibbs free energy of natural and pretreated clinoptilolite were evaluated, and the spontaneity of adsorption is seen to follow the series Na + -form > natural-form > K + -form > Ca 2+ -form.In all cases, the free energy (ΔG°) of the Ni 2+ sorption was negative, suggesting that the spontaneity of the process increased with increasing temperature.The values of the ΔG° also confirm that the maximum adsorption is obtained with the Na +exchanged clinoptilolite followed by the natural, K + and Ca 2+ forms.
The ΔH° was positive in all forms of clinoptilolite and ranged from 18.72 kJ/mol to 42.05 kJ/mol (Table 1), which indicates that the sorption reaction is endothermic for Ni 2+ .These values were calculated from plots of ln K eq versus 1/T.The linear nature of the plot indicates that the mechanism of adsorption is not changed as temperature is changed.But the amount of adsorption is changed because the supply of thermal energy is different.The endothermic nature of the adsorption processes shows that these processes are not energetically stable. 30lf the values of ΔH° for Ni 2+ adsorption had been within the range of 8.4-12.6 kJ/mol, then one could propose that the adsorption process was ionic in nature. 15owever, the values obtained in this study were greater than 12.6 kJ/ mol, which indicates that the mechanism for the adsorption of these ions in zeolites is not ion exchange.

The Ni 2+ -Na + system
The ion-exchange isotherms shown in Figure 1 demonstrate the extreme selectivity of Na + -clinoptilolite for the incoming Ni 2+ cations, which is confirmed by the high values of the equilibrium constant (K eq >1) in Table 3 and the corresponding negative values of the free energies of exchange.The data also indicate that the value of K eq increases with increasing temperature from 298.15 K to 348.15 K in all forms of the zeolite.
The enthalpy of exchange is positive for the Ni 2+ -Na + clinoptilolite.This positive value is explained on the one hand by the substitution of Na + cations by Ni 2+ cations, which have a greater heat of hydration (-406 kJ/ mol for Na + and -2105 kJ/mol for Ni 2+ ) 32 in the aqueous phase, and on the other hand by the greater interaction energy of the Ni 2+ cations with the exchange centres of clinoptilolite as a result of their similarity in size.

The K + -Ni 2+ and natural-Ni 2+ systems
The K + -Ni 2+ and natural-Ni 2+ systems presented highly selective convex isotherms when Ni 2+ was adsorbed; this finding is confirmed by the high values of the exchange constant (~K eq >16 at 323.15 K).The enthalpies of exchange for both of these systems were positive and were almost of the same magnitude (18.75 kJ/mol for K + -Ni 2+ and 22.38 kJ/mol for natural-Ni 2+ ).This result suggests that K + is common to both systems as an exchangeable ion.The Ca 2+ -Ni 2+ system Distinct from the other systems, the Ni 2+ -Ca 2+ system yielded a concave isotherm, clear non-selectivity and the equilibrium constants (K eq ) were less than 1 at 298.15 K and 305.15K (Table 3).The Gibbs free energy of exchange was positive at these temperatures, but a further increase in temperature resulted in the spontaneity of the reaction.The ΔG° values were positive at 298.15 K and 308.15K (3.098 kJ/mol and 0.527 kJ/ mol, respectively) and positive at temperatures greater than 323.15K. Pabalan 22 also reported a positive ΔG° value of 4.19 kJ/mol at 298.15 K for a system involving Ca 2+ -Na + using American clinoptilolite.
The substitution of the doubly charged Ca 2+ cations with doubly charged Ni 2+ ions proved to be an endothermic exchange process (positive enthalpy).The entropy for the Ca 2+ -Ni 2+ exchange system was positive because in the aqueous phase the strongly hydrated doubly charged cations (Ca 2+ ) were substituted by the similarly hydrated Ni 2+ cations.

Conclusion
Evaluation of the thermodynamic parameters K eq , ΔG°, ΔH° and ΔS° provided insight into the mechanism of Ni 2+ sorption by the zeolite.The results of this research showed that clinoptilolite is an effective low cost adsorbent for the removal of Ni 2+ from aqueous solution.Pretreatment of the zeolite enhances the removal efficiency if monovalent ions are used, and selectivity depends on the type of the exchangeable ions on the zeolite.Treating the zeolite with CaCl 2 decreased the zeolite's ability to remove Ni 2+ from aqueous solution.Ca 2+ bind more strongly to the zeolite than sodium and potassium ions.The thermodynamic parameters revealed that Ni 2+ sorption in clinoptilolite is spontaneous and endothermic.The surface area and pore volume data presented for the Na + and Ca 2+ forms were within experimental error, while the K + form had a higher surface area and pore volume.

4 VolumeFigure 1 :Figure 2 :
Figure 1: The Na + -Ni 2+ isotherm at 298.15 K. X(Ni 2+ in soln): equivalent fraction of ingoing nickel in the liquid phase.X(Ni 2+ in Zeo): equivalent fraction of ingoing cation in the solid phase.Percentage error bars for all points are indicated.The solid curve is a polynomial fit of order 2.

Figure 3 :Figure 4 :
Figure 3: The natural-Ni 2+ isotherm at 298.15 K. X(Ni 2+ in soln): equivalent fraction of ingoing nickel in the liquid phase.X(Ni 2+ in Zeo): equivalent fraction of ingoing cation in the solid phase.Percentage error bars for all points are indicated.The curve was fitted to the isotherm data using a polynomial of order 2.
account for the errors caused by instability of the isothermal bath temperature and the calibration of the AAS.
223.15 K, 333.15K or 348.15K for 24 h.2After equilibrium was established, the clinoptilolite was separated from the solution by centrifugation and the pH of the supernatant was adjusted to 7. The equilibrium concentration of the exchangeable ions (Na + , K + , Ca 2+ ) and Ni 2+ in the samples was determined by atomic absorption spectrometry (AAS, Varian 55B).In order to calculate the experimental error, all adsorption experiments were performed in triplicate, which enabled us South African Journal of Science http://www.sajs.co.za to

Table 1 :
Chemical composition (wt %) of treated clinoptilolite samples determined by X-ray flourescenceThe elemental analysis of the natural and pretreated clinoptilolite revealed that it is mainly composed of SiO 2 , Al 2 O 3 and Fe 2 O 3 with very low amounts of MnO and TiO 2 in the framework.The extra-framework ions Na + , K + , Mg 2+ and Ca 2+ showed considerable variation depending on the pretreatment agent used and this change was mainly at the expense of Na + and Ca 2+ content.SiO 2 ranged from 76.13% to 77.37% and Al 2 O 3 ranged from 12.93% to 13.30%.

Table 2 :
Structural characteristics of different homoionic forms of the zeolite 37in solution.The method used to calculate ion activities is the one proposed by Debye and Huckel37in 1923 in which the ionic strength of the solution is calculated first and then the activity coefficient.Because the activity coefficient approaches unity at very low concentrations, Equation 8 can be re-written as 5 Volume 110 | Number 5/6 May/June 2014 South African Journal of Science http://www.sajs.co.zaNi