Stability of Metal Uncharged Ligand Complexes in Ion Exchangers Part 2.-The Copper + Ethylenediamine Complex in Montmorillonite and Sulphonic Acid Resin BY ANDRE MAES, PAUL PEIGNEUR AND ADRIEN CREMES* Centrum voor Oppervlaktescheikunde en Colloi'dale Scheikunde, Katholieke Universiteit Leuven, de Croylaan 42, B-3030 Heverlee, Belgium Received 14th February, 1977 The thermodynamic stability of the copper + ethylenediamine complex is determined in montmoril- Ionite and a macroreticular sulphonic acid resin using ion exchange data and the complex formation function. Both methods are in excellent agreement and correspond to an increase in the overall stability constant of lo3 for montmorillonite and for the resin. The stepwise formation constants FIE2 of the ion exchanged complex are for montmorillonite and and logm7 for the resin, as compared with values of 10'0.7 and logs3 for the bulk solution.Some tentative interpretation is offered in terms of a decreased ligand solvation in the ion exchanger. and In Part 1 it was shown that metal + uncharged ligand complexes in ion exchangers can be characterized thermodynamically by two different methods: a new ion exchange method which leads to the overall stability constant and the extension of the Bjerrum complex formation function, which specifies the stability in terms of the stepwise formation constants of the intermediate complexes. It is the purpose of this paper to illustrate these methods through a thermodynamic study of the copper + ethylenediamine complex in an inorganic and an organic ion exchanger.The bulk of the experimental work in the area of adsorption of metal uncharged ligand complexes is confined to organic ion exchangers (see Part 1) and very few data are available for inorganic ion exchangers. Moreover, it appears that in some cases such as montrnorill~nite,~-~ mechanisms other than ion exchange may be involved, judging from the fact that such complexes are adsorbed in amounts far greater than the cation exchange capacity. Most of the data are therefore of a semi- quantitative nature in that little effort has been, or could be, made to characterize these adsorbed species thermodynamically. Owing to its very high stability, the copper + ethylenediamine complex is well suited for such a quantitative study: first, the fully coordinated two-complex is formed in a very small excess of the ligand, a condition which eliminates any significant ion exchange involvement of the protonated ligand; secondly, it is possible to carry out the distribution measurements of both the aqueous copper ion and the complex at identical, fairly low pH values, a condition which, in the case of montmorillonite with a pH-dependent charge, is most advantageous. EXPERIMENTAL Two ion exchangers are used in this study : the strong acid macroporous 100-200 mesh sodium lewatite SP 1080 (Merck, analytical grade) and the <0.5 pm fraction of the sodium form of the Camp Berteau montmorillonite, prepared as described ear lie^.^ These materials 182A .MAES, P . PEIGNEUR AND A . CREMERS 183 are converted into the calcium form by equilibrium dialysis with 0.5 rnol dm-3 calcium nitrate.The resin is washed free of electrolyte and air-dried prior to use, its water content being measured over P205. The montmorillonite suspensions, adjusted to a clay content of 10 g dm-3 are equilibrated with 0.005 mol dmd3 Ca(N0,)2, prior to use. The cation exchange capacity of the resin is 4.2meqg-I dry resin, as measured by isotopic dilution methods, using 45Ca or 22Na. The ion capacity of the clay, as measured by similar methods varies between 1.05 (Na) to 1.15 (Ca) meq g-l at a pH N 6. The calcium-copper ion exchange equilibria are measured at 25°C by shaking overnight known amounts of resin or clay with Ca + Cu nitrate solutions of varying composition at a pH 2~ 5.5 and at constant total concentration (0.005 mol dm-3). The effect of ethylene diamine (en) is studied using similar methods, adjusting the en/Cu molar ratio to 2 and adding an excess of en corresponding to 2.10-4 niol dm-3.The pH in all systems is in the range 6.8-7. These conditions are consistent with the formation of the Cu(en), complex, the stepwise stability constants of which are : log K1 = 10.72 ; log K2 = 9.31.5 Distribu- tion data are obtained by analysing the equilibrium solutions for copper, using (carbon rod) atomic absorption and 45Ca isotopic dilution. The coordination of the copper ion in the ion exchanger is obtained radiometrically from a duplicate experiment, using 14C-labelled en. The formation function of the copper complex in the resin is measured by equilibrating known amounts of Ca-resin with a constant amount of copper (11 meq g-l), the en/Cu molar ratio varying between 2 and 0.4.The pH values vary in the range of FZ 5 (low molar ratio) to 7 (high molar ratio). A different procedure is used in the case of montmorillonite. 10 cm3 of Na-clay suspension, dialytically equilibrated with 0.01 rnol dnr3 NaN03 are added to 40 cm3 of a solution of sodium + copper nitrate at constant total normality (0.01) containing 1.2 meq copperlg clay. The en/Cu molar ratio is varied from 0.3 to 1.9 and the pH of all systems is adjusted to 7. All systems are shaken overnight and centrifugated. Equilibrium solutions are analysed for copper and en (radiometrically) and their pH measured, using an extended scale potentiometer. dool spacings for the montmorillonite saturated with Cu(en), are measured on an air-dry film, using a Philips PN 1051 diffractometer (CuK,, 3, = 1.5418 A).Heats of exchange are measured at 25°C with the LKB-10700-2 batch microcalorimeterY6 using the mono- protonated enH+-montmorillonite as starting material. The enHf clay is prepared by dialysing Na-montmorillonite with a 0.1 rnol dm-3 en solution at pH = 10.3 and asubsequent series of equilibrations with a 0.05 mol dm-3 en solution at pH = 11.5. Heats of exchange are obtained for the reactions enH+-Cu(en)z+ and enHf-Cs+ and the term for the hypo- thetical reaction Cu2+-Cu(en)$ i- is obtained indirectly from the Na+-Cu2+ and Na+-Cs+ data.6* RESULTS ION EXCHANGE DISTRIBUTIONS The experimental ion exchange isotherms of the aqueous copper ion and the copper + ethylenediamine complex in Ca-montmorillonite and Ca-lewatite are shown in fig.1. In order to afford a quantitative illustration of the effect of en on the ion exchange behaviour of the copper (complex) ion, the data have been represented as the equivalent fraction of copper in the ion exchanger as a function of the logarithm of its equivalent fraction in the equilibrium solution. It was established from the en distribution data that the copper species involved is in fact the Cu(en);+ complex: the average ligand number of copper in the ion exchanger varies between 2 and 2.05. In assigning all of the adsorbed ligand to the copper ions, it is assumed that the ion exchange adsorption of the protonated forms enH+ and enH$+ is occurring to a negligible extent.Such an assumption is easily justified in view of the fact that the equilibria were carried out at 0.005 mol dm-3 total concentration and in the presence of a very low excess of en (2 x mol dm-3). Under the pH conditions used, such an excess would correspond to a concentration of enH+ and enH$+ of fi: mol dm-3, which is nearly two orders of magnitude184 STABILITY OF COMPLEXES IN ION EXCHANGE 4 3 2 1 0 - log SC" FIG. 1.-Ion exchange isotherms in montmorillonite and lewatite for Ca2+-Cu2+ (0, A) and Ca2+-Cu(en)~'(0, A). The dashed curve refers to unit selectivity coefficient : the lower curve refers to lewatite. 7 6 5 4 3 2 I 0 0 . 2 0.4 0.6 0.8 ZC" FIG. 2.-Logarithm of selectivity coefficient against exchanger composition in montmorillonite and lewatite for Caz+-Cu2+ (0, and Ca'+-Cu(en);+ (0, A); the lower curve refers to lewatite.A .MAES, P . PEIGNEUR AND A . CREMERS 185 below the calcium concentration in the equilibrium solution. This assumption is further verified by two experimental facts: first, the adsorption of en in calcium montmorillonite, measured under similar conditions as for the ion exchange isotherms but in the absence of copper, amounts to 1 to 2 % of the exchange capacity ; secondly, throughout the composition range of the exchanger, the sum of adsorbed copper (complex) and calcium ions equals the exchange capacity, pointing to a one-to-one stoichiometry. In montmorillonite, these values vary in the range 1.10-1.1 5 meq g-l and for lewatite, 4.10-4.35 meq g-l. Therefore, we are confident that no mechanisms, other than ion exchange, are involved in the process.TABLE THERMODYNAMIC DATA FOR THE Ca-CU AND Ca-CU(en)z EQUILIBRIA equilibrium In K AGo/kJ mol-1 mon t morilloni t e 0.25 - 0.63 lewatite -0.60 +1.50 7.00 -17.15 2.50 - 6.27 { ZCa+ Cu sZCu+Ca Z C ~ + ~u(en), + ~ ~ u ( e n ) ~ + ~a Fig. 1 shows that for both ion exchangers, there is only a very small difference in ion exchange affinity between calcium and copper. In montmorillonite, Cu is slightly preferred, while the opposite is true for lewatite. The addition of en has a significant effect on copper adsorption: in lewatite, this effect corresponds to a one-order-of-magnitude lowering of the copper concentration in the equilibrium solution, whereas in montmorillonite the shift amounts to three orders of magnitude.All selectivity data are summarized in fig. 2, which shows the variation of the logarithm of the selectivity coefficient against ion exchanger composition. In view of the low total concentration, we may identify the molar ratios with activity ratios. are summarized in table 1, along with the standard free energy values. These results correspond to In K values of 6.75 (montmorillonite) and 3.1 (lewatite) for the reversible displacement of the aqueous copper ion by the Cu(en);+ complex. If we assume a unit partition coefficient for the uncharged ligand, then these data correspond to stability enhancements of - 16.8 (montmorillonite) and - 7.8 (lewatite) kJ mol-l, i.e. the overall stability constant of the Cu(en)$+ complex is raised by factors of The thermodynamic equilibrium constants, as obtained by graphical integration and respectively.COMPLEX FORMATION FUNCTIONS The complex formation data for the Cu(en)$+ complex in both exchangers are represented in fig. 3. In the case of montmorillonite, the data correspond to copper loadings 2,- in the range 0.85-0.95, these values referring to the lower and higher ligand concentrations. For lewatite, these loadings vary in the range 0.3 to 0.9. The free ligand concentration in the equilibrium solution is calculated from pH values, copper concentration and overall ligand content using the formation constants for the copper complex, reported in the experimental section and log K,(enH+) = 10.17 ; log K,,(enH;') = 7.49.5 The ligand numbers E of copper in the ion exchanger are net values, i.e.the ion exchange contribution from the protonated forms (predominantly enH5-t) is sub- tracted from the overall ligand content. This correction was estimated using an upper limit of K, = 10 for the Ca-enH;+ equilibrium in the resin and K, = 30 for the Na-enH$+ equilibrium in the clay. This correction is quite small, at most 0.05 in 5, as can be expected on the basis of the very low concentrations of enH$f, relative186 STABILITY OF COMPLEXES I N I O N EXCHANGE to Na+ and Ca2+ ions (enH;+ M mol dm-3 for the clay) and the fact that enH2f has to compete with either Na+ or Ca2+ for a small fraction of the ion capacity. Fig. 3 shows that complex formation is initiated at much lower ligand concentra- tion, compared with the bulk solution, the effect being most pronounced in mont- morillonite.The curves in fig. 2 afford a direct estimate of the average formation constant, defined as K1K2, from the reciprocal of the free L concentration, corres- ponding to a degree of complex formation of 0.5; the result is for mont- morillonite and for lewatite. mol dm-3 for the resin and M 20 1.6 12 - n 0.8 0.4 - - - - - - -l 8 9 I 0 r i 12 PL FIG. 3.-Ligand number of the copper ion in montmorillonite (0) and lewatite (0) against logarithm of free ligand concentrations in solution. The dashed curve refers to complex formation in solution, The stepwise stability constants were obtained by least-squaring the data in the equation - K, L + 2K,R,L2 n = I+K,L+R,R,L~ and the resulting data are summarized in table 2.The overall stabilization factor, obtained by this method is lO3*I and 101.32 for montmorillonite and lewatite, a result which is in excellent agreement with the result obtained from ion exchange measurements. TABLE ~.-STABLLITY CONSTANTS FOR THE Cu(en), COMPLEX, OBTAINED FROM COMPLEX FORMATION DATA overall stabilization IogF, logZ2 log82 factor montmorillonite 11.60 11.50 23.10 10 3.1 lewatite 11.65 9.70 21.35 bulk solution 10.72 9.31 20.035 10 1.32 X-RAY AND CALORIMETRIC DATA The 001 spacing of Cu(en)2-montmorillonite, obtained as the average of the first, third and fourth-order reflections is 12.6 AA. MAES, P. PEIGNEUR AND A. CRBMBRS 187 The microcalorimetric results are summarized in fig. 4. The ion exchange displacement of enH+ by Cu(en):+ is slightly endothermic, whereas the reaction with Cs+ is considerably exothermic.Using literature values for the AH terms for Na+-Cs ’ and Na+-Cu2+ exchange, we estimate the heat terms for the hypothetical reaction Z Cu,, + Cu(en):+ + Cu(en)$+ + Cu&+ to be about - 12 kJ mol-l. cnH Ns I i I I 1 I I f+ 18.0 t t f I I I I I I I I 1 I cs FIG. 4.-Heats of exchange reactions (kJ/equivalent) in montmonionite. DISCUSSION The foregoing data show that the Cu(en)$+ complex is strongly stabilized in both ion exchangers. The stabilization corresponds to a factor of loo0 (montmorillonite) and 20 (lewatite) in the overall stability constant, assuming a partition coefficient of unity for the ligand. The result found for the resin is in excellent agreement with the value which may be calculated from the data of Cockerel1 and Walton on the same complex (logp, N 21.5) and with the data reported by SkoroWood and Kalinina lo (log Kl = 11.39 to 11.77 and log K2 = 9.74-9.83, depending on moss linking).The nature of the agreement tends to show that the stabilization is insensitive to the degree of cross linking, since our data were obtained on a highly porous resin, whereas the literature data refer to resins of 10-16 % DVB. The good agreement between the ion exchange method and the complex formation indicates that the reaction is thermodynamically reversible. It should be emphasized that the extent of stabilization, as obtained from ion exchange, refers to the actual stability constant in the equilibrium solution and is therefore independent of its absolute value.A similar situation occurs in the complex formation method where the ratio of stability constants in the ion exchanger and the solution f12/b2 (or K l / K ,188 STABILITY OF COMPLEXES I N ION EXCHANGE and K2/K2) is quite insensitive to the absolute values of the bulk solution stability constant, chosen for calculating the free ligand concentration : for example, a 10 % difference in Kl and K2 has an insignificant effect on the extent of stabilization obtained. No unambiguous interpretation has been given for the stabilization of such complexes in ion exchangers. Cockerel1 and Walton proposed that the diamine acts as a " cross linking ligand " binding metal ions in a chain or network. Sometimes the " organic solvent character " differences of ion hydration and lower dielectric constant in the ion exchanger 1 1 * l2 are thought to be responsible for complex stabilization.In montmorillonite, the X-ray data, which agree with literature data,l 3-1 point to the presence of a 3 A monolayer between the lamellae and it appears that Cu(en)$+ is present as a monomeric species: its conformation is essentially identical to the one in aqueous solution, i.e. a square planar complex in which the two axially coordinated water molecules are replaced by oxygens from the clay lattice. Such a configuration was also proposed for the hydrated copper ion under conditions where a monolayer is present in the interlamellar region and where the copper ion is thought to be coordinated to four water molecules in X Y plane.It appears therefore that the cross-linking hypothesis is not tenable, at least not for mont- morillonite, since monodentate ligands such as thi~urea,~ pyridine and alkylamines, to be discussed in subsequent papers, are found to lead to equally important stabiliza- tion phenomena in montmorillonite. Invoking the organic solvent character as an explanation does not seem realistic either, in view of the much larger effects in montmorillonite. If, of course, the observations were limited to organic resins, it could be argued that the stabilization is merely an artefact resulting from a very favourable distribution of the ligand between the ion exchanger and the liquid phase, causing the adsorbed metal ion to be " exposed " to higher ligand concentrations than those measured in solution.However, the less pronounced effect in resins fails to support this idea. In any case, the fact that the metal complex exhibits a very high preference for the ion exchanger under conditions (high ligand concentration) corresponding to the formation of the same coordinatively saturated complex in both cases, clearly points to a thermo- dynamically real effect. It is likely that the reason for the enhanced stability of ion exchanged complexes is related to the effect of the ion exchanger on solvent properties and that the more important effects in montmorillonite are connected with surface geometry. Of course, bulk properties such as dielectric constant can be invoked, but it seems preferable to search on the molecular level. It is apparent that the origin of the stabilization in montmorillonite is mainly energetic in origin.The experimental value of - 12 kJ mol-1 is subject to some uncertainty, being determined indirectly. However, unpublished data from this laboratory, obtained for the same complex on hectorite from temperature dependence of ion exchange equilibria show a similar trend. That the origin was predominantly energetic was also shown from diffuse reflectance spectroscopy on dehydrated Cu(en)$+ montmorillonite and interpreted in terms of an increase in crystal field stabilization energy. Turning our attention to the solvent effect, it is customary to analyse differences in solvent properties in the two phases exclusively in terms of possible effects on cation hydration. In view of arguments presented above, it seems that the primary hydration of the copper ion in montmorillonite is hardly affected, apart from the replacement of the axially coordinated water molecules by lattice oxygens.In the present contextA . MAES, P . PEIGNEUR AND A . CREMERS 189 however, it is important to examine possible effects of a change in solvent properties on the ligand. In a series of papers connected with the enhancement of thermodynamic stability of metal complexes with tetramine macrocyclic ligands, compared with the non- cyclic analogues, Margerum and co-workers 17-19 introduced the concept of a " macrocyclic effect ". It was shown that the stability enhancement was mainly due to a more favourable enthalpy term which could not be described in terms of stronger metal-nitrogen bonds in the cyclic ligands.The enthalpic differences were attributed to a decreased ligand solvation of the cyclic ligand which has less amine hydrogen bonded water to be displaced in the complex formation reaction. Such ligand solvation effects " will hold for any ligands where the donor groups are forced to be close to one another or in some way are shielded from solvation ".18 This situation could occur in the shallow interlamellar region in montmorillonite where hydrogen bonded solvent-lattice structures have been demonstrated.20 Admittedly, these are little more than conjectures, the value of which remains to be further tested, perhaps through a thorough thermodynamic study of a number of complexes in various ion exchangers.The financial support of the Belgian Government (Programmatie van het Wetenschapsbeleid) is acknowledged. A. Maes, P. Marynen and A. Cremers, J.C.S. Faraday I, 1977, 73, 1297. S. L. Schwartzen Allen and E. Matijevic, J. Colloid Interface Sci., 1975, 50, 143. M. H. El-Sayed, R. G. Burau and K. L. Babcock, Soil Sci. Soc. Amer. Proc., 1971, 35, 571. J. Pleysier and A. C. Cremers, J.C.S. Furaday I, 1975, 71, 256. G. L. Sill6n and A. E. Martell, Stabllity Constants of Metal Ion Complexes (The Chemical Society, London, 1964-1971). ti A. Maes, P. Peigneur and A. Cremers, Proc. Int. Clay Conf. (Applied Publishers, 1975), p. 319. A. Cremers and H. C. Thomas, Israel J. Chem., 1968, 6, 949. G. L. Gaines and H. C. Thomas, J. Chem. Phys., 1953,21,714. L. Cockerell and H. F. Walton, J. Phys. Chem., 1962, 66, 75. lo 0. R. Skorokhod and A. A. Kalinina, Russ. J. Phys. Chem., 1975,49,187. l1 0. R. Skorokhod and A. G. Varawa, Russ. J. Phys. Chem., 1972,46,980. l2 Y. Marcus and A. J. Kertes, Ion Exchange and Solvent Extraction of Metal Complexes (Wiley, l 3 W. Bodenheimer, L. Heller, B. Kirson and S. Yariv, Clay Min. Bull., 1962, 5, 145. 14R. Laura and P. Cloos, Proc. Rkunion Hispano-BeIga de Minerales de la Arcilla (Madrid, l5 F. Velghe, R. A Schoonheydt, J. B. Uytterhoeven, P. Peigneur and J. H. Lunsford, J. Phys. l6 D M. Clementz, T. J. Pinnavaia and M. M. Mortland, J. Phys. Chem., 1973,77, 195. l7 D. K. Cabbiness and D. W. Margerum, J. Amer. Chem. SOC., 1969,91,6540. '* F. P. Hinz and D. W. Margerum, J. Amer. Chem. SOC., 1974,96,4993. l9 F. P. Hinze and D. W. Margerum, Inorg. Chem., 1974,13,2941. 2o V. C. Farmer and J. D. Russell, Trans. Faraday SOC., 1971,67,2737. London, 1969). 1970), p. 76. Chem., submitted for publication. (PAPER 7/243)