CHAPTER 1
Electrolyte Solutions
BY A. K. COVINGTON AND T. H. LILLEY
PART I: Thermodynamic Properties of Electrolyte
Solutions 1 Introduction
In the period under review water has remained the most studied solvent system, although more thermodynamic studies are being made using the newer protic and aprotic solvents. A major preoccupation has been with ion-solvent interactions and particularly with 'solvent structure effects', a loose phrase which, as Atkinson has said, may be no more than a collective repository for our ignorance whilst we are still forced to retain a continuum theory of ion-ion interactions. In the following sections it will be noted that tetra-alkylammonium salts have been extensively studied by nearly all available techniques and much discussion has revolved around interactions between large tetra-alkyl-ammonium ions and the solvent. Whilst we are still a long way from a complete understanding of the factors involved, progress is being made particularly by modern spectrnscopic techniques and for this reason we devote a separate section entirely to discussion of this topic.
In order to improve their knowledge of non-ionic solution interactions a number of 'electrolyte' physical chemists have turned their attention to aqueous non-electrolyte solutions. Kineticists too, have become more concerned with attempts to understand the effect of environmental factors on ionic reactions. The proceedings of a symposium held in Newcastle in January 1968 which brought together those with a common interest in hydrogen-bonded solvents, have been published, as have lectures presented at an earlier meeting held in Bradford. 'Solute-Solvent Interactions' is the title of a collection of essays centred around, rather than on, this theme. A review of hydration effects and the thermodynamic properties of ions has recently appeared. A review essentially complementary to the present one has appeared in Annual Reports on the Progress of Chemistry for 1968.
'Structure and Properties of Water' is the title of a book by Eisenburg and Kauzmann and of a review by Ives and Lemon. Views on this subject remain subjective and often controversial. The translation of the proceedings of a conference held in Tbilisi in 1966 on 'Water in Biological Systems' has been published and contains articles by Gurikov on 'Water Structure' and by Samoilov on the 'Theory of Hydration in Aqueous Solutions'. Franks has reviewed the role of water structure in disperse systems. A computer-compiled bibliography of papers covering the period 1957 — 68 on water structure and the physical properties of water is available 16 from Bell Telephone Laboratories, Murray Hill, New Jersey. Anomalous, meta or polywater so-called, has aroused considerable interest even though it was first reported by Deryagin in 1962. It is confidently expected to be the subject of a number of papers in 1970 whilst arguments continue about its observed physical properties and structure.
The viscosity of (ordinary) water up to 1400 kg cm-2 and moderate temperatures (2 — 30 °C) has been reported. Recent data are considered to be in error. Kay and co-workers have traced discrepancies in reported values for the dielectric constant (ε) of water at various temperatures to a doubtful capacitance correction. The preferred values are given by the equation:
log ε = 1.94409 - 1.991 x 10-3 T
A minimum has been reported 27 in the Kerr constants for H2O and D2O near 30 °C, which doubtless is related to structural features. Recalculations of the heat capacity of water at constant volume have been given which are important for testing models of liquid water. Energies of proton solvation have been derived from studies of the generation of protonated water molecules in a high pressure gaseous ion source.
Two other solvents of considerable current interest, dimethyl sulphoxide (DMSO) and dimethylsulphone, have been studied by neutron inelastic scattering and by X rays. Similarity of liquid and solid spectra shows a high degree of dipole association. The effect of these solvents on water structure was also studied. X-Ray studies are also reported of formamide and of potassium iodide solutions in formamide. It was concluded that concentrated solutions show ion-pair formation and doubly solvated cations.
Physical properties of solvents reviewed, or newly reported, include dimethyl sulphoxide, N-methyl acetamide and its mixtures with dioxan and benzene (dielectric constants), N-methyl propionamide (density and dielectric constants, 20 — 40 °C) and sulpholan (tetramethyl sulphone) and its mixtures with water and methanol (density, viscosity, and dielectric constant).
Gillespie has reviewed his recent work with the highly acidic solvent, fluorosulphuric acid, in which such interesting ionic species as I2+ (red), I42-, Se82+ (green), Se42+ (yellow) and Te42+ (red) have been shown to exist. A successful synthesis of perbromates has been reported by electrolytic oxidation of bromate, and rubidium perbromate has been isolated from the chemical oxidation of bromate with xenon difluoride. Zordan and Hepler present a critical Latimer-type compilation of data for manganese and its ions.
2 Single Electrolytes
A. Activity and Osmotic Coefficients. — Cation responsive glass electrodes (see Chapter 2) are now sufficiently well established as reliable electrodes that, with care, they can be used for precise determination of activity coefficients. Such measurements are always made relative to some chosen standard concentration in the same way as determination of activity coefficients from cells with transport and transport numbers. Hostetler, Truesdell, and Christ report activity coefficients for potassium chloride in the range 10 — 50 °C using potassium-sensitive glass electrodes. A rather complicated graphical method of evaluation was devised to eliminate the effects of e.m.f. drifts with time. Truesdell has reported similar determinations for sodium chloride, where such elaborate analysis is unnecessary because of the better behaviour of the sodium-responsive electrode. Because of interference from hydrogen ions, the pH of the solutions is important. Schindler and Waelti, using a small quantity of added 'tris' to buffer the solution, have confirmed accepted values for sodium chloride activity coefficients in the range 0.13 — 2.2 molal. A paper by Shatkay and Lerman also reports some measurements on sodium chloride solutions using sodium selective and silver-silver chloride electrodes. Activity coefficients of ammonium nitrate in liquid ammonia at -30 °C were obtained from hydrogen electrode concentration cell measurements and previously determined transport numbers.
Salomon reports activity coefficients from e.m.f. work for lithium chloride and bromide in propylene carbonate and lithium bromide in anhydrous DMSO. Lithium metal and amalgamated thallium-thallium halide electrodes were used. Butler and co-workers have determined activity coefficients for lithium chloride in anhydrous DMSO from e.m.f. measurements and also studied the same system by cryoscopy. Their results, however, are at variance with those of Garnsey and Prue, and Dunnett and Gasser. It is suggested that systematic errors are present in the latter two investigations, possibly due to heat transfer effects giving rise to spurious steady temperatures for periods up to thirty minutes. This view is contentious but the differences do not arise from use of the wrong cryoscopic constant. λc = 4.07 kg mol-1 K was determined from measurements with benzoic acid. Whereas the value used by Dunnett and Gasser was higher (4.36), their raw results of freezing point depression are up to 0.02° higher than those of Garnsey and Prue. The latter workers also report cryoscopic measurements on alkali-metal perchlorates and lithium chloride in sulpholan (tetrahydrothiophene-1,1-dioxide). The advantages of the large cryoscopic constant (λc = 64.1 ± 0.2 kg mol-1 K) are largely illusory for thermal buffering is poor and differences in the solidus and liquidus slopes are small. DMSO and sulpholan are solvents of different powers of ionic solvation even though they have the same dielectric constant, with the result that the opposite sequence of osmotic coefficients with cationic radius for the alkali-metal perchlorates is observed. These perchlorates are incompletely dissociated in sulpholan and there are large differences in the association behaviour of lithium chloride and bromide in this solvent. Another paper reviews the properties of sulpholan as a solvent for cryoscopy.
Russian workers reports some new determinations and some recalculations of osmotic coefficients from cryoscopic measurements for alkali-metal halides, chlorates, and bromates in aqueous solution. Bonner, Kim, and Torress have carried out cryoscopic studies of various solutes in ethylene carbonate and N-methyl acetamide (NMA). They note that the order of osmotic coefficients for the alkali-metal iodides (Li > Na > K > Rb > Cs) is the same for these two solvents as in water at any fixed concentration. Further the osmotic coefficient of bis-trimethylammonium iodide is nearly the same as in water. Since this can be considered as a dimer of tetramethylammonium iodide they conclude there is no structure enforced ion-pairing in aqueous solutions of tetra-alkylammonium salts.
Comparative (isopiestic) vapour pressure measurements are more popular than direct measurements but a few such determinations have been reported. Gardners has revised and extended previous work on the osmotic co-efficients of sodium chloride at high temperatures (125 — 270 °C). Previous data at 1 molal were low by 2% for reasons not ascertained. Using a differential manometric method, Bus, Steinberg, and de Boer have determined vap2ur pressures in the system D2SO4-D2O, and Campbell and Oliver in the systems lithium or sodium chlorate-dioxan-water. It was concluded that dioxan plays a major role in ionic solvation. By a gas displacement method the partial pressures of hydrogen chloride over saturated lithium chloride (13 molal) containing hydrochloric acid were determined indicating appreciable incomplete dissociation.
A useful discussion of important features for the design of isopiestic apparatus has been given by Luk'yanov, who describes in detail apparatus in which equilibrium attainment is enhanced by the device of rotating the container vessel on a central pivot at an angle of ca. 20° to the vertical. A method 60 of double isopiestic measurement where two volatile components are employed, is somewhat restricted in its application because of solubility restrictions.
A modified isopiestic apparatus employing glass flasks instead of the more usual silver or platinum dishes has been described by Wai and Yates and used to determine the water activity of concentrated perchloric acid solutions extending the range of study beyond 16 molal (62 — 75%).
Two papers report osmotic coefficients for bivalent perchlorates. Libus and Sadowska have studied manganese, cobalt, nickel, and copper perchlorates comparing their results with literature data for zinc and magnesium. Over the range 1.0 — 3.5 molal the osmotic coefficients of cupric and magnesium perchlorates are slightly lower than those of the other four. Pan and Ni report values for cadmium perchlorate and chloride. Platford has used the isopiestic method to study sodium and potassium tetraborates, sodium metaborate and fiuoroborate, and boric acid. The latter behaves like a non-electrolyte up to saturation. An apparatus suitable for studies up to 80 °C has been used to obtain data for potassium chloride and bromide and sodium sulphate, which can be compared with boiling-point elevation studies at reduced pressures (60 °C). Alkylsulphonium salts which are similar in behaviour to alkylammonium salts have been studied by Lindenbaum.
A number of papers report the use of vapour pressure osmometers to determine osmotic coefficients. Your reporters share the views of others who consider the accuracy of this method is often exaggerated especially if water is the solvent. Sometimes called the adiabatic isopiestic method, it depends on observation of the condensation of solvent on to droplets of solution and solvent placed separately on two identical thermistors.
Schwabe, Kretschmar, and Gartner report partition measurements of uranyl perchlorate between water and some organic media from which they derive activity coefficients for uranyl perchlorate much lower (3 orders of magnitude) than those determined from isopiestic measurements by Robinson and Lim. The new measurements are also supported by ultracentrifuge studies. Whilst the explanation of the abnormally high values determined isopiestically, may lie in hydrolysis, this also affects the new measurements. Since the actual experimental data are not given, it is not possible to comment further.
From distribution measurements of carboxylic acids and their sodium salts between water and butyl ether, Czeisler and Schrier have derived activity coefficients for the acids and salts in the aqueous phase. Dimer formation of the acids is assumed in the organic phase.
B. Volumes. — Determination of apparent and partial molal volumes from precision density measurements continues to attract considerable attention. In an important series of papers, Dunn describes a precise dilatometric technique and reports determinations of densities of sodium chloride, potassium chloride, bromide, and iodide and also tetra-n-butyl ammonium bromide at 25 °C in the range 0.001 — 1.0M. The third paper extends measurements to seven temperatures in the range 0 — 65 °C confirming the Debye-Hückel slope for volume at these temperatures for barium and calcium chlorides in addition to the alkali-metal salts mentioned above. It was noted that φov (apparent molal volume at zero concentration) increases with temperature reaching maximum values at 60 °C for the 1:1 electrolytes, 35 °C for calcium chloride and 45 °C for barium chloride. The effect is attributed to changes in the water structure on temperature increase. Structural effects are often invoked to explain any anomalous behaviour that is encountered. An entertaining paper by Holtzer and Emerson should be compulsory pre-reading for all those tempted to commit their fancies to print. Desnoyers and co-workers, and Millero and Drost-Hansen report precision determinations by float methods for 1:1 electrolytes in water. The latter paper gives data for 0.1 molal solutions at 1° intervals in the range 20 — 40 °C. A cubic fit to this density data is differentiated 79 to yield apparent molal expansibilities. The same authors with Korson find no evidence for reported anomalies in the temperature dependence of the apparent molal volume (or viscosity) of sodium sulphate near 32·5 °C. Ellis has continued his work on densities of electrolyte solutions at high temperatures (up to 200 °C).
Apparent molal volumes of aqueous solutions of sodium alkyl sulphates (C10 — C14) have been reported by Franks and co-workers. φv exhibits negative deviations from the limiting law, both above and below the c.m.c. at which there is a large positive volume change. Possible explanations of the first have been summarised by Franks and Smith who prefer the 'co-operative hydrophobic hydration effects' explanation. Prue and Pethybridge clearly distinguish the various schools of thought on the topics of 'hydrophobic bonding' and 'hydrophobic interaction' which are current in connection with the many recent studies of tetra-alkylammonium salts over the past few years. Other opinions on these topics can be found in the proceedings of a symposium on 'Hydrogen-Bonded Solvent Systems'. Desnoyers and Arel find that plots of φv - 1.86 c+ versus c are linear for RNH3Br where R = H to octyl, which they take as evidence that dimerisation cannot explain a minimum and subsequent increase in φv for R ≥ C7. Millero and Drost-Hansen report measurements of φv for R4HCl where R = H to n-butyl at 1° intervals between 20 — 40 °C at one concentration only. Broadwater and Evans find that the behaviour of octane-1,8-bi-(tri-n-butylammonium)dibromide is similar to that of n-butylammonium bromide. The former is considered as a model for a cation-cation pair which is another possible explanation of observed 'hydrophobic hydration' effects. Conway and Laliberté have determined the isotope effect for alkali metal and tetraalkylammonium salts in light and heavy water. A positive effect V2°(D2O) > V2°(H2O) is found for structure-making ions (tetra-alkylammonium) and negative for structure breaking ions (sodium fluoride). Since the sizes of H2O and D2O are identical it is considered that the differences must reflect the degree to which the solutes affect the structure of the solvent, and therefore D2O is more structured than H2O. Darnell and Greyson using a dilatometer technique have investigated the effects of solutes on the temperature of maximum density of D2O(tmax 11.17 °C cf. 3.98 °C for H2O). tmax is reduced by all salts in both isotopic waters by an extent proportional to the solute concentration. Lithium chloride reduces tmax in both solvents and therefore is considered to be a structure breaker at this temperature (but not at 25 °C). With the exception of this salt, lowering is greater in H2O than in D2O, which might suggest that more structure is broken in H2O than D2O in spite of the fact that the latter is usually considered to be more structured. The preferred explanation for the observations that structure making effects are not found at tmax is unconvincing.
