J . Chem. SOC., Faraday Trans. 1, 1982, 78, 2041-2049 Thermodynamic Study of the Influence of Complexation on Exchange Equilibria in Wyoming Bentonite Clay BY ANDRE MAES,* EDDY RASQUIN AND ADRIEN CREMERS Centrum voor Oppervlaktescheikunde en Collo'idale Scheikunde, Katholieke Universiteit Leuven, De Croylaan 42, B-3030 Leuven (Heverlee), Belgium Received 9th February, 198 I In Wyoming Bentonite the ion-exchange reactions for the homovalent pairs Ag(en);-Cs+ and Cu(en)i+-Ca2+ and for the heterovalent pair Ag(en)$-Cu(en)z+ are characterized by A G e values of exchange which are very much larger than those for the aqueous metal ions. These effects, which emphasize the importance of ion-ligand uersus ion-solvent (water) interactions, are discussed in terms of the relative contributions of the interactions occurring in the surface and bulk solution phase. A significant decrease is found in bivalent ion selectivity with increasing loading in the Ag(en)l-Cu(en)i+ case (two orders of magnitude difference in K , at both ends of the composition scale).This is opposite to what is commonly found in mono-divalent ion-exchange equilibria in montmorillonite. This effect is discussed in terms of heterogeneity in charge distribution in the exchanger and variations in electrostatic interaction. The exchange selectivities of the aqueous transition-metal ions Zn2+, Ni2+, Cu2+, Co2+ and Cd2+ versus Na+ in montmorillonite are quite similar,l the overall AG* value varying around - 0.75 kJ equiv.-l. In contrast, ethylenediamine complexes of Zn2+, Ni2+, Cu2+ and Ca2+ are very selectively exchanged,' the overall AG* values for the Ca2+ displacement ranging from - 7 (Zn2+) to - 10.25 (Ca2+) kJ equiv.-l.The present paper reports on the exchange of two complexes, Ag(en)l (versus Cs+) and Cu(en)i+ (versus Ca2+) and their aqueous counterparts in Wyoming Bentonite (W.B.) clay. The purpose is twofold: (1) to gather more information on the role of the ion-solvent interaction on the magnitude of the exchange selectivity by modi- fying the ion environment using a complexation with neutral ligands; (2) to verify whether the characteristic selectivity rise with bivalent ion loading which is observed in heterovalent exchange reactions among hydrated cations3* is also observed in heterovalent exchanges among complexes. EXPERIMENTAL The fraction of Wyoming Bentoniteclay smaller than 0.5 pm was separated usingconventional techniques, and was stored as a 3% Na+ suspension in 1 mol dm-3 NaNO, at 5 OC in the dark.Appropriate quantities of W.B. clay were dialysed to 0.01 mol dmP3 Na+ prior to use. The Cu(ethylenediamine)i+-Ca2+ selectivity was measured by overnight equilibrium of 10 cm3 of a 1 % Ca2+ suspension of W.B. with 20 cm3 of mixed Cu(en)i+-Ca2+ solution of 0.01 total normality in Cu2+ + Ca2+. The formation of the fully coordinated Cu(en)i+ complex was ensured by adding an excess of 2 x mol dm-3 ethylenediamine (en). The pH of the starting solutions was brought to 8.5 with HNO,. The Ag(en),f-Cs+ equilibrium in W.B. was measured by mixing 5 cm3 Na+ suspension of the clay (1 %) at 0.01 mol dm-3 NaNO, with 20 cm3 mixed solutions of Ag(en),f and Cs+ at 0.01 total normality.From the stability constants of the Ag(en)z complex (Kl = lo4.', K2 = 103.0)5 we calculated that a free ligand concentration of 0.1 mol dm-, (en) at pH 11 was necessary to form the biscomplex. The solutions were therefore 0.125 mol dmP3 in (en) and at pH ca. 11.5. 204 12042 EXCHANGE EQUILIBRIA IN WYOMING BENTONITE The selectivity between the Ag(en)l and Cu(en)i+ complexes was measured at 25 and 5 OC by equilibrating 10 cm3 1 % Na+ suspension of W.B. with 40 cm3 of mixed Cu(en);+-Ag(en)$ solutions of 0.0 1 total normality in Cu2+ + Ag+. The (en) concentration was 0.125 mol dmP3 and the pH was ca. 11.5. The binary Na+-Ag+ and Na+-Cs+ exchange reactions were also studied at 0.01 total normality using an Na+ suspension of W.B.at 0.01 mol dm-, NaNO, and mixed Na+-Ag+ and Na+-Cs+ solutions, respectively. All equilibrations were performed in a thermostatted environment at 25 OC by overnight end-over-end shaking, followed by phase separation using a thermostatted supercentrifuge. Concentrations were determined by radiotracer methods using 22Na, lloAgm and 13'Cs. Cu2+ was analysed by flameless and conventional atomic absorption techniques (varian Techtron AAG). d(OO1) spacings of wet, fully exchanged Cu(en)i+ at 5 x loP4 and 0.1 mol dm-, free (en) and Ag(en)l at 0.1 mol dmP3 free (en) were measured using the Debye-Scherrer technique. RESULTS The natural logarithm of the selectivity coefficients for the Cs+-Ag(en)$, Ca2+-Cu(en)i+ and Cu(en)i+-Ag(en)$ equilibria in W.B.are shown in fig. 1, 2 and 3 as a function of the equivalent fraction of Ag(en)i, Cu(en)i+ and Ag(en)z, respectively. ZAg FIG. 1 .-Dependence of In K,[Cs+-Ag(en)z] on silver loading. In K,(Cs+-Ag+), shown by the broken line, is calculated (see text). The sum of Ca2+ +Cu(en)g+ adsorbed ranges from 0.84 to 1 .O mequiv. g-l at the Ca2+- and Cu(en)i+-rich ends of the isotherm. The equilibrium pH is 7. In the case of the Cs+-Ag(en)l equilibrium the sum total ranges from 0.96 mequiv . g-l (Cs+-rich end) to 1.08 mequiv. g-' [Ag(en)l-rich end], and the equilibrium pH is 1 1.5. The use of a well-dispersed Na+ clay as starting material in the Cs+-Ag(en)$ and Cu(en)i+-Ag(en)l equilibria is preferred over using the flocculated Ag(en)i, Cs+ or Cu(en);+ clays because the former allows easier handling and better reproducibility in pipetting.The levels of Na+ displaced in the liquid phase are sufficiently lowA. MAES, E. RASQUIN A N D A. CREMERS FIG. 2.-Dependence of In Kc[Ca2+-Cu(en),2+] on copper loading. In Kc(Ca2+-Cu2+), shown by the broken line. is taken from the literature.s 2043 I ' I I I I I 1 I I 0.2 0.4 0.6 0.8 ZA, ZAg FIG. 3.-Dependence of In K,[Cu(en)i+-Ag(en);] on silver loading at 5OC (A) and 25OC (0). In K,(Cu2+-Ag+), shown by the broken line, is calculated (see text). Corrections for solution-phase activity ,.-dX,.:-..tA n,-o :.,,-.l-.A-A bUGlIIblG1lLJ a l G 1 1 1 ~ 1 u u G u . (ca. 4 x I 0-3 mol to justify an analysis in terms of a purely binary system. Indeed, using K , (Na+-Cs+) = 10.25 (see later) one estimates the maximum Na+ occupancy to be ca.5% of the Cs+ content. The amount of Na+ adsorbed therefore equals 0.05 mequiv. g-l at high Cs+ contents and vanishes towards the Ag(en)i-rich end. In the case of the Cu(en)i+-Ag(en)i exchange, the selectivities of both cations versus Na+ are high enough to displace Na+ from the exchange complex entirely.2044 EXCHANGE EQUILIBRIA I N WYOMING BENTONITE Z A g FIG. 4.-Dependence of In K,:(Na+-Ag+) on silver loading. I I I I I 1 3 t 4 ZCS FIG. 5.-Dependence of In K,(Na+-Cs+) on caesium loading. Kielland plots for the Na+-Ag+ and Na+-Cs+ exchanges are given in fig. 4 and 5. The total ion occupancy, i.e. Ag+ + Na+, remains constant at 0.85 f 0.02 mequiv. g-l. Beyond an equivalent fraction of Ag+ of 0.7-0.8 an oxidoreduction phenomenon occurs, as evidenced by the blackening of the sample.The selectivities in this region are therefore extrapolations. The sum of Na+ + Cs+ increases gradually from the Na+ CEC of 0.82 mequiv. g-l to 0.98 mequiv. g-l at the Cs+-rich end. All selectivity data were obtained at low total normality (0.01), which allows us to identify the activity ratio with the concentration ratio in solution in the cases of homovalent exchange, i.e. the Cs+-Ag(en)l, Ca2+-Cu(en)i+, Na+-Ag+ and Na+-Cs+ exchanges. In the case of the heterovalent Ag(en)i-Cu(en)i+ exchange the activity coefficient ratio equals y"g(en)l - rf Ag(en), NO, rCu(en>f+ 7% Cu(en),(NO,), which may differ from unity. The Davies equation6 fits the activity coefficients of all electrolytes up to I = 0.2 and was used to calculate the solution-phase activity coefficient ratio.The ionic strength was obtained by accounting for Na+, Cu(en):+, Ag(en)i and the monoprotonated form of (en). The correction, expressed in terms of In K, units, then varies from 0.266 to 0.249 over the experimental composition range and is included in the data in fig. 3. The selectivities for the exchange reactions of aqueous ions were determined or taken from literature data and are represented by the broken lines in fig. 1-3. In K(Cs+-Ag+) = -2.05 and is calculated from In K(Na+-Cs+) = 2.45 andA. MAES, E. RASQUIN A N D A. CREMERS 2045 In K(Na+-Ag+) = 0.4, which were obtained by graphical integration of the experi- mental data in fig. 4 and 5; In K(Ca2+-Cu2+) = - 0.044 is as given by El Sayed et af.;8 In K(Ag+-Cu2+) = - 0.784 is obtained from the former Ca2+-Cu2+ and Ag+-Na+ data and In K(Na+-Ca2+) = 0.06;9 In K,(Ag+-Cu2+) is then -0.784+ 1 = 0.216. All thermodynamic data, obtained by using the Gaines and Thomas integrati~n,~ are given in fig. 6 in terms of AGe of exchange and are combined with literature data. I \ \ A I E! d h 0 d I Ag+ *-0.986 N a+ \ c s+ FIG. 6.-Combination of thermodynamic data given in kJ equiv.-' of the indicated binary equilibria in Wyoming Bentonite clay. Full and broken lines represent experimentally obtained and calculated exchanges, respectively. Literature data are indicated where used. TABLE l.-AG* VALUES (kJ mol-l) OBTAINED AT 25 OC FOR THE INDICATED HYPOTHETICAL The overall stability constants corresponding to the formation of Cu(en)i+ and Ag(en)l in W.B.v2) and in the solution phase ( p2) are also compared. EXCHANGES Cu2+-Cu(en)i+ - 14.41 22.57 20.03 Ag+-Ag(en); - 16.40 10.59 7.7 AGe(Na+-Cs+) = -6.04 kJ equiv.-l differs markedly from the value obtained by Gast et a(.', in another bentonite, but fits perfectly into the charge-density relationship for the Na+-Cs+ equilibrium in octahedrally substituted montmorillonites.ll Fig. 1 and 2 clearly demonstrate that complexing of Ag and Cu with (en) results in high selectivities for both cations. In K[Ca2+-Cu(en)i+] = 5.80, whereas In K(Ca2+-Cu2+) is only -0.044. These data fit into the charge-density relationship found in other montmori1lonites.l2 Complexing of Ag+ enhances its selectivity uersus Cs+ from In K(Cs+-Ag+) = - 2.05 to In K[Cs+-Ag(en)z] = 4.60.Complexation has a similarly dramatic effect on exchanges among complexes, as seen in fig. 3 ; the In K value for the Ag(en)i-Cu(en)i+ is -7.84 and 0.47 for the aqueous ions. The free-energy changes for the hypothetical exchange reactions of the uncomplexed by the complexed ions are easily calculated from these data and are given in table 1. They correspond to the extra stabilities gained by the complexes upon adsorption in the interface.13 These free-energy changes were shown13-14 to be related to the overall complex formation constants in the surface g2) and solution phase v2) by Pz2046 EXCHANGE EQUILIBRIA I N WYOMING BENTONITE in which M and MLn stand for the aqueous metal ion and its coordinatively saturated complex.The data in table 1 show an increase in the logarithm of the overall stability constant of Ag(en); and Cu(en);+ complexes in W.B. by, respectively, 2.89 and 2.54 units with respect to their value in bulk solution. The foregoing data are useful in assessing the reversibility of the Cu(en)i+-Ag(en)t exchange reaction. From the cycle in fig. 6 one calculates a free-energy change of - 10.17 kJ equiv.--l in favour of the Ag(en)l complex. This is in good agreement with the experimentally obtained value (- 9.67 kJ equiv.-l), proving the reversibility of exchange reactions of complexed ion species. The good correspondence between calculated and experimental AG*[Ag(en)i- Cu(en)X+] values is remarkable in view of the fact that equilibria involving Ag(en)t were obtained at high pH, whereas the Na+-Ca2+, Na+-Cs+ and Ca2+- Cu(en)t+ equilibria were studied at neutral pH values.The formation of a Cu(en)i+ complex at high (en) concentration and high pH in solution and/or at the surface might invalidate the use of the Cu(en)i+-Ca2+ equilibrium at pH 7 in the cycle (fig. 6) to calculate the Ag(en)i-Cu(en)t+ equilibrium. However, the formation of Cu(en)i+ does not occur in solution under the present experimental conditions. Indeed, the U.V. spectrum of a solution containing 0.25 mol dmP3 Cu2+ and 1 mol dm-3 (en), which should be sufficient to make the triscomplex if K3 were 1 ,5 only shows the 18 000 cm-l band of the biscomplex Cu(en)i+. The observed d(OO1) spacing of 1.34 nm on Cu(en)t+-W.B. suspension at 0.1 mol dm-3 (en) and pH 11.5 also points to the formation of a biscomplex (see table 2).In Camp Berteau Montmorillonite, Cu(en)i+ species can be generated under anhydrous conditions from Cu2+-C.B. and gaseous (en). However, on contact with air, Cu(en)i+ is immediately transformed into Cu(en)i+. l5 The involvement of Cu(en)i+ in W.B. is therefore believed to be minimal. TABLE 2.--d(OO 1) SPACING (nm) OF FULLY EXCHANGED Ag(en)l AND Cu(en),"+ WYOMING BENTONITE CLAY AT LOW AND HIGH FREE (en) CONCENTRATION 2 x 10-4 no complex formed 1.5 1 x 10-1 1.34 1.34 DISCUSSION 'THERMODYNAMICS OF EXCHANGE OF COMPLEXES A characteristic feature of ion exchange involving complexes in montmorillonites is the overall increase of AG* of exchange, as compared with the value for the aqueous counterparts.Important changes occur in AG* when one of the exchanging ions is complexed with a ligand such as (en), as exemplified in the pairs Cu(en)i+-Ca2+, Ag(en)t-Cs+ and the Cu(en)i+-Cu2+ and Ag(en)t-Ag+ systems (see fig. 6 and table 1). Considering exchanges among complex ions one finds AG* changes of ca. 2 kJ equiv.-l for the exchanges of the (en) complexes of Ni2+ and Zn2+, Ni2+ and Cd2+ in Camp Berteau Montmorillonite.2 The selectivity among the aqueous ions, however, is very sma1l.l AG* [Ag(en):-Cu(en)i+] is similarly high (9.67 kJ equiv.-l) compared with AG* (Ag+-Cu2+) of the aqueous ions (0.967 kJ equiv.-l). TheseA. MAES, E. RASQUIN A N D A. CREMERS 2047 observations stress the importance of the ion-ligand and/or ion-water entities in determining the magnitude of the selectivity. The present results corroborate the observations made for alkali-metal16 and alkaline-earth17 ion exchanges in montmorillonite, in the sense that homovalent exchange equilibria involving complexes in montmorillonite are also exergonic when the most polarizable cation [Ag(en)i or Cu(en)i+] exchanges for a less polarizable one (Cs+ or Ca2+).This relation holds in cases where the polarizability differences are high (complex uersus aqueous ions) or for homologous series of ions as the alkali metals. Any change in the thermodynamic state functions of exchange relates to a difference of such functions, for both ions concerned, at the surface and in the solution phase. In terms of the Eisenmann18 model, exchanges in montmorillonite follow the low field-strength pattern, i.e.the sign of the AG, AH and TAS terms is governed by the solution properties of the ions. In the case of alkali-metal and alkaline-earth ion exchange the sign of all three thermodynamic functions is predicted. Those exchange reactions are always exergonic, exothermic and occur with a negative entropy change when an ion is replaced by one with smaller AG, AH and A S terms of hydration. The parallel with the polarizability parameter is self-evident. The influence of complexation on exchange becomes apparent by comparing the thermodynamic functions in the presence and absence of (en). Taking the Cu(en)i+-Ca2+ case as an example, the equations take the form - AG,en) = (Gcucen), - G d - (Gcucen,, - Gca) in which G stands for the free-energy content.The bar denotes the surface phase. The change in the overall AG of exchange on complexing corresponds in sign to the variation in the solution terms. Indeed, smaller AG values of hydration of the complex compared with the uncomplexed cation are expected. In the case of exchange among complexes the following relevant equations are written for the Ag+-Cu2+ system in the presence and absence of (en): AG = (Ccu - 2 G A g ) - (GCu - 2G,,) AG(en) = (Gcuccn), - 2-dAg(en),) - (Gcu(en), - 2G,g(en),)* Here again the variation in the overall AG* on complexing parallels in sign the variation in the solution term. The exact magnitude of the changes in hydration energy on complexing may be expected to be high, but are difficult to assess, and consequently it is not possible to deduce the separate contributions of both surface and solution terms to the overall AGe change.The importance of the surface term can be judged from the surface charge-density dependence of the Ca2+-Cu(en)i+ and Ca2+-Cu2+ exchange equilibria in a series of montmorillonites of variable charge density.12 AG(Ca2+-Cu2+) is almost independent of charge density, whereas AG[Ca2+-Cu(en)i+] decreases linearly from ca. - 10 kJ equiv.-l at 1.5 x mequiv. cm-2 to zero at vanishing charge density. (The charge in W.B. corresponds to 1 x lop7 mequiv. cmP2.) CHANGES I N SELECTIVITY WITH SURFACE COMPOSITION Kc[Ag(en),'-Cu(en)i+] decreases by more than two orders of magnitude with increasing bivalent ion loading (see fig. 3), in contrast to the commonly observed opposite behaviour in heterovalent exchange reactions among hydrated cati0ns.l.3, Two alternatives have been proposed to explain the general observation of increasing K,(Mn-M1+n) with M1+n ion loading in the case of heterovalent exchanges2048 EXCHANGE EQUILIBRIA IN WYOMING BENTONITE among aqueous metal ions in montmorillonites. First, the increase in K,(Na+-Ca2+) with Ca2+ content is explained as an autocatalytic effect provoked by the progressive collapse of double layers as exchange proceeds and by the fact that the selectivity for planar sites exceeds that for edge sites.19 Secondly, configurational entropy terms arising from the replacement of a monovalent by a divalent cation have been invoked3 to explain the overall entropy gain, because of the larger number of possibilities of arraying bivalent versus monovalent cations and also to explain the K,(M+-M2+) increase with bivalent ion loading.Heterovalent exchange among the complexes considered occurs in collapsed layers 4 A thick, as indicated by the d(OO1) spacings of the end members (table 2), in contrast to the more swollen state in the Na+-Ca2+ case. The selectivity changes with loading can therefore not be explained by a progressive interlayer collapse. Using a hypothetical charging-discharging process it has been predicted20 that the overall free energy of heterovalent exchange reactions increases with charge-density increase. This was verified for the Na+-Ca2+ exchange in montmorillonites of variable charge density.21 The mean charge density of W.B.clay itself is a composite of different charge densities.22 One may therefore expect to observe the highest affinity for the Cu(en)$+ ion at the small bivalent ion loadings corresponding to the exchange occurring in the highly charged regions of the clay. This is experimentally verified by the observation that the relative selectivity for Cu(en)i+ is highest at the low Cu(en)i+ constant side of the Kielland plot in fig. 3. The observed changes in Cu(en)i+-Ag(en)i selectivity with loading could then merely result from Coulombic interaction terms, and are thus of energetic origin. This is verified by two observations. (1) From the temperature dependence in fig. 3 it is calculated that the overall Ag(en)l-Cu(en)i+ exchange reaction itself is governed by enthalpic factors (AH x 12.5 kJ equiv.-l).( 2 ) Unpublished results on the charge-density dependence of the Ag(en)l - Cu(en)i+ exchange agree with the predicted increase in AG*[Ag(en)i-Cu(en)i+] with charge- density increase. The conclusion is therefore that configurational entropy terms may contribute to the same extent to the selectivity changes with loading in both heterovalent exchanges between complexed and uncomplexed cations. However, in the case of exchange among complexes energetic terms exceed the entropic contribution and are responsible for the observed selectivity changes with loading. We thank the Belgian Government (Programmatie van het Wetenschapsbeleid), the Fonds voor Kollektief Fundamenteel Onderzoek (F.K.F.O.) and the K. U. Leuven (3e cyclus Fonds) for financial support.A. Maes, P. Peigneur and A. Cremers, Proc. Znt. Cfay Con$ (Applied Publ. Ltd, 1976), p. 319. P. Peigneur, A. Maes and A. Cremers, Proc. Znt. Cfay Con$ (Elsevier, Amsterdam, 1979), p. 207. M. B. McBride, Clays Cfay Min., 1980, 28, 255. R. Van Bladel, G. Gaviria and H. Laudelout, Proc. 4th Inf. Cfay Con$ (Pergamon Press, Oxford, 1972), vol. 2. p. 15. G. L. Sillen and A. E. Martell, Stability Constants of Metal-ion Complexes (The Chemical Society, London, 1964-1971). C. N. Davies, Zon Association (Butterworths, London, 1962). G. L. Gaines and H. C. Thomas, J . Chem. Phys., 1953, 21, 714. M. M. El Sayed, R. G. Burau and K. L. Babcock, Soil Sci. SOC. Am., Proc., 1970, 34, 397. J. Dufey, Ph.D. Thesis (Universite Catholique de Louvain, 1974). lo R. G. Gast, R. Van Bladel and K. B. Deshpande, Soil Sci. SOC. Am., Proc., 1969, 33, 661. l1 A. Maes and A. Cremers, J . Chem. SOC., Faraday Trans. 1, 1978, 74, 1234. l2 A. Maes and A. Cremers, J . Chem. SOC., Faraday Trans. I , 1979, 75, 513. l 3 A. Maes, P. Marynen and A. Cremers, J . Chem. SOC., Faraday Trans. I , 1977, 73, 1297. l4 A. Maes, P. Peigneur and A. Cremers, J . Chem. SOC., Faraday Trans. I , 1978, 74, 182.A. MAES, E. R A S Q U I N A N D A. CREMERS 2049 l5 F. Velghe, R. A. Schoonheydt, J. B. Uytterhoeven, P. Peigneur and J. H. Lunsford, J . Phys. Chem., l6 H. Martin and H. Laudelout, J . Chim. Phys., 1963, 60, 1086. l7 H. Laudelout, R. Van Bladel, G. H. Bolt and A. L. Page, Trans. Faraday SOC., 1968, 64, 1477. l8 G. Eisenmann, Biophys. J., 1968, 2, 259. lS I. Shainberg and W. D. Kemper, Clays Clay Min., 1966, 14, 117. *O G. H. Bolt and C. J. Winkelmolen, Zsr. J. Chem., 1968, 6, 175. *l A. Maes and A. Cremers, J . Chem. SOC., Faraday Trans. I , 1977, 73, 1807. 22 M. S. Stul and W. J. Mortier, Clays Clay Min., 1974, 22, 391. 1977, 81, 1187. (PAPER 1 /211)