J. Chem. SOC., Faraday Trans. I , 1982, 78, 43-52 Solute-Solvent Interactions in Water + t-Butyl Alcohol Mixtures Part 12.-Single-ion Enthalpies of Transfer using the Tetraphenylarsonium Te trap hen yl bora te Assump ti on BY JEAN JUILLARD Laboratoire d’Etude des Interactions Solutes-Solvants, Universite de Clermont 11, B.P. 45, 63170 Aubiere, France Received 29th September, 1980 Heats of solution of tetraphenylarsonium chloride (4,AsCl) and sodium tetraphenylboride (NaB4,) in water and water+t-butyl alcohol mixtures (from 0 to 40% by weight) have been measured. From these data and the previous values for NaCl, enthalpies of transfer of tetraphenylarsonium tetraphenylboride are estimated. Using the classical assumption of equivalent change of solvation with solvent media of the two tetra-aryl ions, ionic enthalpies of transfer are estimated for various cations and anions.This reveals a completely distinct behaviour for the two types of ions. These results are discussed in terms of the structure of the water + t-butyl alcohol mixtures. In previous papers we have reported data for some thermodynamic properties of transfer of various solutes from water to an exemplary series of water + organic media, the water + t-butyl alcohol mixtures. The interpretation of the experimental results is made difficult because each species involved in the process of solvation cannot be considered separately. Therefore it was decided to carry out a splitting into ionic contributions using various extra thermodynamic assumptions. A review of the various processes used in such a splitting into individual ionic transfer and solvation properties has been recently provided by C0nway.l Various methods can be used: (1) Extrapolation of salt values, with respect to ionic radius or ionic volume (or mass), to zero ion size or zero reciprocal radius.Such a method used data for homologous series of salts having either a common anion or a common cation. (2) Use of the assumption that an ion and a neutral molecule similar in size, shape and surface have the same properties of transfer. (3) Splitting the value for a salt composed of ions of similar radius and shape into its ionic component values. Such a method is considered to be reliable for large and chemically similar cations and anions in a salt, e.g. d4P+B4; or b4As+B&.As far as the first type of splitting method is considered, many types of extrapolation were applied to the present problem using both alkali metal halides and tetra- alkylammonium halides. None of them corresponds to linear plots or gives consistent results. In the second type of method some hope was placed in the idea that the enthalpies of transfer of both the cryptates of alkali ions and the corresponding free ligands could be considered as identical. However, it has been shown recently by Abraham,2 who studied enthalpies of transfer from water to methanol in such systems, that this assumption is not relevant. Finally a method of the third type, which is now considered to give more reliable results, was used : data concerning enthalpies of transfer of tetraphenylarsonium 4344 INTERACTIONS I N WATER + t-BUOH MIXTURES tetraphenylboride (B#,#,As+) and their use for obtaining ionic enthalpies of transfer for numerous cations and anions in water+TBA mixtures are reported here.Tetraphenylboride tetraphenylphosphonium could have been considered as well but it has been proved in some cases that the two assumptions are eq~ivalent.~ Such separations using tetra-aryl ions have been criticised by Treiner, on the basis of ' scaled-particle theory' calculations of the Gibbs free energy of transfer. However, as emphasised by Abraham,5 Treiner's criticisms result from a misunderstanding of the nature of this assumption. Some data were already available on enthalpies of transfer of tetraphenylarsonium tetraphenylboride (#4A~B$4) in water + TBA media.These enthalpies were obtained from measurements of the heat of precipitation of this electrolyte in various mixtures by Bright and Jezorek.6 They can also be calculated using data reported by Arnett'? for sodium tetraphenylborate (NaB4,) and tetraphenylarsonium chloride (4,AsCl). Preliminary calculations made from these two sets of data give such surprising results for the enthalpies of transfer of anions and cations that it appears necessary to confirm these results before discussing further what can be deduced from them. Therefore, in order to check and extend these results to mixtures with higher alcohol concentrations and to specify the location of the maximum, we measured the heat of solution of NaB4, and #,AsCl in water+TBA mixtures from 0 to 40% by weight. From the additivity of ionic contributions to the enthalpies of transfer, AHt of #,AsB#, can be obtained and the ionic enthalpy of transfer can thus be estimated for various cations and anions from the AH, values of electrolytes previously reported.9-16 EXPERIMENTAL PART Sodium tetraphenylborate and tetraphenylarsonium chloride were pure commercial products (Fluka puriss and pro analysis) used as received. Purification of solvents and calorimetric measurement procedures have already been described.l0 RESULTS Results obtained for the heats of solution of NaB#, and #,AsCl are given in table 1. The standard enthalpies of solution were obtained using the Debye-Huckel calculation of the heat of dilution as previously de~cribed.~ Fig.1 shows clearly that TABLE 1 .-ENTHALPIES OF SOLUTION OF SODIUM TETRAPHENY LBORIDE AND TETRAPHENYLARSONIUM CHLORIDE FROM WATER TO WATER+t-BUTYL ALCOHOL MIXTURES AT 25 "c XU Xb NaBq4," q3,AsCI" 0 5 10 15 20 25 30 35 40 0.0000 0.01 26 0.0263 0.041 1 0.0573 0.0749 0.0943 0.1157 0.1394 - 18.59 8.82 -8.10 21.64 +6.16 36.97 f41.51 50.40 + 49.34 47.96 + 17.03 37.18 + 9.24 - + 4.40 34.15 + 30.53 - a Weight % of alcohol in the mixture; mole fraction of alcohol in the mixture; AH? in kJ mo1-I (mean uncertainty k0.05).J . JUILLARD 45 A 90 80 - 70 - h 60- . 5 0 - - h 24 3 40 - 30 - 20 - FIG. 1.-Enthalpies of transfer from water to waterfTBA mixtures (mole fraction x of TBA) of: (l), tetraphenylarsonium chloride (right-hand scale): A, our values; A, corrected values of Arnett’ (see text) ; (2) sodium tetraphenylborate (left-hand scale); a, our values; 0, from Arnett and McKelvey;* (3) tetraphenylarsonium tetraphenylborate (left-hand scale) : 0 , calculated from our measurements [eqn (I)] ; 0, from heat of precipitation (Bright and Jezorek); 0, calculated from Arnett’s data [eqn (2)].data obtained by previous authors61 are roughly in agreement with ours. Values of enthalpies of transfer of tetraphenylarsonium tetraphenylboride from water to water+TBA mixtures have thus been confirmed and specified. For NaB4, there is a good agreement between our values and those of Arnett, as shown in fig. 1 , in which both sets of enthalpies of transfer are compared. For #,AsCl, Arnett’s values estimated from fig. 1 of ref. (7) and corrected for the transfer of H,017 are slightly lower than ours. Owing to the uncertainty of such a procedure, the discrepancies observed can be considered as acceptable.Enthalpies of transfer of #,AsB4, calculated from our data and eqn (1) are compared in fig. 1 with the value obtained by Bright and Jezorek6 from enthalpies of precipitation and also with the values calculated using eqn (2) and the data of Arnett and McKelvey7~* as proposed by Holterman and Eng1e~ts.l~ Enthalpies of transfer of NaCl and H,O used in the calculations are taken, respectively, from ref. (9) and (16): AHt(#,Asq5,B) = AH,(Na#,B) + AHt(q5,AsCl) - AH,(NaCl) (1)46 INTERACTIONS I N WATER + t-BUOH MIXTURES AHt(4,As4,B) = AH,(Nad,B) + AHt(4,AsC1, 2H,O) - 2AHt(H,0) - AH,(NaCl). (2) A mean curve was drawn using both Jezorek’s data and ours, and the enthalpies of transfer of the two ions were calculated from eqn (3): Another procedure could have been used (for example, considering only our data), but in any case a separation of ionic contributions should not give exact values but only show any trend in variations of the ionic enthalpies of transfer.From the C1- and Na+ contributions thus obtained, data of ref. (9) give the enthalpies of transfer of K+, Rb+ and Cs+, data of ref. (12) give H+ and Li+, and data of ref. (14) give Br-, I- and C10;. The values for NO; are obtained from Pointud’s unpublished results for AH, of KNO,. The enthalpies of transfer of OH- are obtained from the values previously calculated for H+ and from the dissociation enthalpy and partial molal enthalpy of water in water + TBA media.16 AHt for substituted benzoates are deduced from data reported in ref.(13) and AHt for acetate, propionate and isobutyrate are obtained from AH, of acetic, propionic and isobutyric acidslo combined with the changes in the enthalpies of dissociation of these acids.ll Finally, ammonium and tetra-alkylammonium ionic enthalpies of transfer are obtained from AH, of their halides reported in ref. (15). All these results are presented in fig. 2-5. There are no other determinations of individual ionic enthalpies of transfer from water to water+ TBA mixtures in the literature. However, some arguments confirming the trend in the variations observed here can be based on recent determinations by Gilletl8 -15 - 10 - 5 - 0 0.05 0.10 FIG.2.-Enthalpies of transfer of some inorganic monovalent cations from water to water + TBA mixtures (mole fraction x of TBA): dashed line, Na+; full lines, (1) H+, (2) NH: (left-hand scale), (3) Li+, (4) Cs+, (5) Rb+, (6) K+ (right-hand scale).J. JUILLARD 47 5 t 0.05 0.10 X FIG. 3.-Enthalpies of transfer of some inorganic monovalent anions: dashed lined, Br-; full lines, ( I ) I-, (2) ClO;, (3) NO;, (4) C1-, (5) OH-. 0.05 0.10 FIG. 4.-Enthalpies of transfer of some carboxylate ions: (1) acetate, (2) propionate, (3) isobutyrate, (4) benzoate, (5) p-iodobenzoate.48 INTER A C T I 0 N S I N WATER + t-BUOH MIXTURES 40 30 20 t o 1 I a 0.05 0.10 FIG. 5.-Enthalpies of transfer of tetra-alkyl and tetra-aryl ions: dashed line, TPB- and TPAs+; full lines, (1) NH:, (2) NMe:, (3) NEt:, (4) NPra, (5) NB:, (6) NPen:.of the enthalpies of acid dissociation of anilinium and pyridinium ions in these mixtures. It is frequently accepted that variations in the solvation of anilines and corresponding anilinium cations with the nature of the solvent media are analogous. Such an assumption is at the root of the use of ' Hammet indicators' in defining acidity functions. This means that changes in enthalpy of dissociation of anilinium would be roughly equal to the enthalpy of transfer of H+. A comparison of enthalpies estimated in this way and enthalpies of transfer obtained here is comforting. Even if the values estimated from treatment of anilinium dissociation data are lower than values obtained here, the same trend can be observed (a maximum in AHt for x z 0.05 but of ca.7 as compared with 15 kJ mol-l). DISCUSSION The results obtained are, on the whole, quite surprising. The most striking fact is the very distinct behaviour on the one hand of inorganic cations and tetra- alkylammonium ions and on the other hand of inorganic anions and even of small organic anions. In all the curves of transfer enthalpies as a function of composition, three parts can be distinguished which probably correspond to zones of distinct structure types for the water + t-butyl alcohol mixtures. Zone 1 : For mole fractions from 0 to 0.03 the enthalpy changes are frequently weak and generally positive, both for cations and anions. This zone is very rich in water; the water structure is frequently described as being enhanced by the presence of the alcohol.Zone 2: For mole fractions from 0.03 to 0.08 notable variations in the enthalpies of transfer can be observed, underlined by the existence of large extrema (maximumJ. JUILLARD 49 or minimum) which are located at an alcohol mole fraction of ca. 0.05; according to the charge, the enthalpy changes with content are very different. There is an endothermic maximum for the cations and an exothermic minimum for the anions. This range of concentration is frequently described as a region where there is a maximum structuring of the mixtures, which corresponds to the formation of an original water + alcohol structure. Zone 3 : For mole fractions from 0.08 to 0.15 there can again be observed a distinct type of behaviour according to the charge of the inorganic ion.When the alcohol concentration increases the enthalpy of transfer becomes more and more exothermic for the cations and more and more endothermic for the anions, except for those which have a hydrophobic radical. In this region there is an increase in free molecules or aggregates of TBA. STRUCTURE OF THE WATER+TBA MIXTURES The structure of water + alcohol and specially of water + t-butyl alcohol mixtures has been discussed by many a u t h o r ~ , l ~ - ~ ~ on the basis of both thermodynamic and spectroscopic data. Alcohols are solutes which bear both hydrophilic and hydrophobic groups. These latter determine their ‘ structure-promoting ’ characteristic which is of the same type as the one observed with apolar solutes.The ‘structure-promoting’ ability of the alcohols increases with the size of the alkyl group. It is generally accepted that when adding alcohol to water there is a reinforcement in the stability of the ‘clusters’. The real type of the structure itself is not yet well-known. However, numerous arguments have been presented for the formation, as in the case of apolar solutes, of a structure analogous to the solid clathrate, but of a more labile type. This was first proposed by Arnettlg to explain the properties of the t-butyl alcohol mixtures and formalised by A t k i n ~ o n ~ ~ in terms of the fluctuating-cage model. The use of this model has been supported by work carried out recently by Iwasaki and F ~ j i y a m a ~ ~ on concentration fluctuations from light-scattering spectra and by Tamura et a1.26 on ultrasonic absorption of the mixtures.The picture of the structure of the mixtures which results is not in clear contradiction to that previously proposed by Symons and Blandamer.21 If one accepts the idea of a two-state mixture model for water it can be said that the introduction of an alcohol molecule will determine a stabilisation of the ‘clusters’ which corresponds to a gradual organisation into fluctuating cages ofwater surrounding an alcohol molecule. These cages have a well-defined stoichiometry. The water structure, at first little affected (zone l), becomes more and more organised when the alcohol concentration is raised, up to a concentration corresponding to the stoichiometry of the cages (zone 2). This stoichiometry, 1 : 17 for a solid clathrate of type 11, is doubtless slightly different for the solution, taking into account the degree of freedom of the cages one with another; the work of Iwasaki et al.25 suggests a value of 1:21.A solid t-butyl alcohol +water clathrate is not known, but a compound where such a clathrate is stabilised by two molecules of H2S for one of TBA and 17 of water has been described.27 Clathrate hydrates of type 1128 present two types of holes, large and small, able to encage molecules, and their structure is stabilised when the small holes are also occupied. Above a certain composition of the mixtures which corresponds to a maximum formation of the cages and thus to a structuration maximum, two phenomena occur: the apparition of free TBA molecules and the condensation of the cages25 thus forming ‘merged clathrates’21 (zone 3).50 INTERACTIONS IN WATER + t-BUOH MIXTURES VARIATION OF THE ENTHALPY OF TRANSFER WITH ALCOHOL CONTENT FOR THE VARIOUS TYPES OF IONS For all the ions studied here, an extremum in AH, (minimum or more frequently maximum) is obtained for TBA contents between mole fractions x = 0.045 (1 : 22) and x = 0.055 (1 : 18).In fact the location of the maximum is not very accurate, taking into account the various calculations involved in estimating ionic enthalpies of transfer; nevertheless such results permit us to discuss our results in terms of the model of TBA + water mixtures presented above (the model of fluctuating cages) even if the stoichiometry of the cages cannot be determined precisely from our data.The results which concern the cations are plotted in fig. 2 (inorganic monovalent cations) and 5 (tetra-alkylammonium ions). For the first results (for x < 0.03) only small changes in the enthalpy of transfer are observed initially and these can be attributed to the persistence of the types of water-cation interactions which exist in water, the cages appearing in the medium having only one effect, that of restricting the quantity of water able to participate in these interactions. The same effect is observed for the first tetra-alkylammonium ions; however, as the size increases, adding t-butyl alcohol to water causes an increasingly rapid variation in the enthalpy of transfer. We underlined in the previous paper (after Desrosiers and Desn~yers~~) that the enthalpy corresponding to the formation of the cavity, according to the size of the ion, is sufficient to explain the magnitude of the effect observed.Whatever the type of cation, a large increase in the enthalpy of transfer is observed above x = 0.03. This enthalpy goes through an endothermic maximum for a composition of the mixture which corresponds roughly to the stoichiometry of the ' pseudo-clathrate'. For the tetra-alkylammonium cations the amplitude of the maximum is a function of the size of the cation. It can be suggested that the introduction of such cations can only be carried out by the (at least partial) destruction of the local structure, this destruction being more and more pronounced as the size of the cation increases.This effect would be strongest for an alcohol content which corresponds to the highest structuring, i.e. to the greatest cage density. For a different reason H+, NH; and the alkali metal cations would also, as a consequence of their ability to attract water molecules, disrupt the structure of the mixtures where most of the water molecules are engaged in the formation of the cages. The sequence of the observed effects for the alkali metal cations, Na+ > K+ > Rb+ > Cs+ > NHZ, is in agreement with the electrostrictive power of these ions (as reflected by the difference between aqueous molar volumes and crystallographic ones30). Compared with this series both H+ and Li+ show a slightly anomalous behaviour. It can be supposed that these small ions can occupy an interstitial location in the structure.Beyond this zone of maximal structuration the enthalpies of transfer decrease more rapidly for the inorganic ions than for the tetra-alkylammonium ions when the alcohol is added. This could be due to interactions between cations and the free TBA molecules which present some basicity as indicated by data from Kolthoff and Chantooni for H+ solvation in alcohols.31 A smaller decrease, due allowance being made in the transfer enthalpy for the tetra-alkylammonium ions, could result from the interference of various processes : a decrease with the size of ions in the interactions related to the basicity of TBA, an increase in the hydrophobic interactions between TBA and the tetra-alkylammonium ions, and competition between TBA and tetra-alkylammonium ions in the building of water clathrates.It is not possible to know the respective participation of these various effects in the enthalpy changes observed.J. JUILLARD 51 If we now consider the inorganic anions (fig. 3) and the small organic anions (fig. 4) the variations in the enthalpies of transfer are analogous to what is observed for cations in the same zone, rich as it is in water (zone 1): there are only small endothermic variations whose weakness is probably related to the lack of large changes in the water-anion interactions in this range of composition. These variations in enthalpy nevertheless appear to increase with the size of the organic ion. This can again be attributed to the increase in the enthalpy of cavity formation with the size of the anion.What happens for anions in zone 2 is more difficult to explain. There is here, for all but the larger anions, an enthalpy minimum which also appears at the ‘magic fraction’. Inversion occurs only for anions as large as bromobenzoate. If we again refer to the fluctuating-cage model, it can be said that anions are more easily accommodated in the clathrate-like mixture than in water but increasingly less so as their size increases. It may be suggested that both inorganic and small organic anions can occupy interstitial locations between the cages where they, like H2S in the solid clathrate, strengthen the structure. Both shape and size would probably determine their ability to fill the vacant holes in the fluctuating structure.Influence of size is clear in the case of small organic ions (fig. 5 ) for which enthalpy of transfer from water to the mixture (x = 0.05) is negative, but increasingly smaller when the size of the ions increase; for inorganic ions there is no clear link between the size and the amplitude of the minimum. Except for the larger ions, the variations in AH, observed in zone 3 (x = 0.08-0.15) are again opposed for anions and cations. The increase in AH, when adding alcohol would be related to the weakness of the interactions between anions and free TBA molecules. The fact that such an effect is smaller for nitrate, perchlorate and benzoate ions is probably related to the development of interactions through dispersion forces between TBA molecules and these anions, in which electron delocalisation is high and which are consequently highly solvated in organic media.CONCLUSIONS To sum up, as far as enthalpies of transfer are concerned, the behaviour of cations and anions, at least of the small ones, is very different. This fact, not always easy to explain, might cause some doubts as to the relevance of the extrathermodynamic assumption used here. But it must be mentioned that the ionic molar volume separation realised by Dollet et al. in the same media,32 this time using a rigorous method developed by Zana and ye age^-,^^ leads to the same sort of observation: there is a decrease in the partial molal volumes of the halide anions and an increase in the partial molal volumes of the alkali cations when going from water to the clathrate-like mixture.In addition separation by Wells3* of ionic contributions to Gibbs free energies of transfer, using other assumptions, leads also to a very different effect for cations and anions: AG, values are positive for anions and negative for cations. Our most recent investigations confirm this It must therefore be accepted that anions and cations are not accommodated in the same way in water+t-butyl alcohol mixtures. Considering the formation of a quasi-clathrate fluctuating structure followed by its destruction allows us, more or less, to explain the effect observed in the change of the solvation of ions, as far as enthalpic effects are concerned. This cannot be considered as proof of the validity of such a structural description of the mixtures, but only as an attempt to fit thermodynamic data on the solvation of ions in these media to what seems at the moment to best reflect the structure and interactions in water + t-butyl alcohol mixtures.J. JUILLARD 51 If we now consider the inorganic anions (fig.3) and the small organic anions (fig. 4) the variations in the enthalpies of transfer are analogous to what is observed for cations in the same zone, rich as it is in water (zone 1): there are only small endothermic variations whose weakness is probably related to the lack of large changes in the water-anion interactions in this range of composition. These variations in enthalpy nevertheless appear to increase with the size of the organic ion. This can again be attributed to the increase in the enthalpy of cavity formation with the size of the anion. What happens for anions in zone 2 is more difficult to explain.There is here, for all but the larger anions, an enthalpy minimum which also appears at the ‘magic fraction’. Inversion occurs only for anions as large as bromobenzoate. If we again refer to the fluctuating-cage model, it can be said that anions are more easily accommodated in the clathrate-like mixture than in water but increasingly less so as their size increases. It may be suggested that both inorganic and small organic anions can occupy interstitial locations between the cages where they, like H2S in the solid clathrate, strengthen the structure. Both shape and size would probably determine their ability to fill the vacant holes in the fluctuating structure.Influence of size is clear in the case of small organic ions (fig. 5 ) for which enthalpy of transfer from water to the mixture (x = 0.05) is negative, but increasingly smaller when the size of the ions increase; for inorganic ions there is no clear link between the size and the amplitude of the minimum. Except for the larger ions, the variations in AH, observed in zone 3 (x = 0.08-0.15) are again opposed for anions and cations. The increase in AH, when adding alcohol would be related to the weakness of the interactions between anions and free TBA molecules. The fact that such an effect is smaller for nitrate, perchlorate and benzoate ions is probably related to the development of interactions through dispersion forces between TBA molecules and these anions, in which electron delocalisation is high and which are consequently highly solvated in organic media. CONCLUSIONS To sum up, as far as enthalpies of transfer are concerned, the behaviour of cations and anions, at least of the small ones, is very different. This fact, not always easy to explain, might cause some doubts as to the relevance of the extrathermodynamic assumption used here. But it must be mentioned that the ionic molar volume separation realised by Dollet et al. in the same media,32 this time using a rigorous method developed by Zana and ye age^-,^^ leads to the same sort of observation: there is a decrease in the partial molal volumes of the halide anions and an increase in the partial molal volumes of the alkali cations when going from water to the clathrate-like mixture. In addition separation by Wells3* of ionic contributions to Gibbs free energies of transfer, using other assumptions, leads also to a very different effect for cations and anions: AG, values are positive for anions and negative for cations. Our most recent investigations confirm this It must therefore be accepted that anions and cations are not accommodated in the same way in water+t-butyl alcohol mixtures. Considering the formation of a quasi-clathrate fluctuating structure followed by its destruction allows us, more or less, to explain the effect observed in the change of the solvation of ions, as far as enthalpic effects are concerned. This cannot be considered as proof of the validity of such a structural description of the mixtures, but only as an attempt to fit thermodynamic data on the solvation of ions in these media to what seems at the moment to best reflect the structure and interactions in water + t-butyl alcohol mixtures.