Acid strength and its dependence upon the nature of the solvent

The development of our views concerning the nature of acids and bases and, in particular, the precision given to these views by the definitions proposed by Brönsted and Lowry of an acid as a substance which splits off protons, and of a base as a substance which takes up protons, have led to a much clearer understanding of the behaviour of acids and bases in different solvents. The essential dependence of the ionization of acids and bases upon the basicity or acidity of the solvent has been emphasized in a number of papers, and many authors have shown how by suitable choice of solvent a much greater range of acidity is available than when water alone is employed. Despite the very notable advances that have been made, there is a further problem, that of the relative strengths of different acids, to which a satisfactory solution has not yet been found. So far as any single solvent is concerned, it is usual to regard the dissociation constant of an acid as a measure of its strength, and on this basis numerous attempts have been made to correlate acid strength and constitution. Many of these attempts have been expressed quantitatively, and that the opinion is widely held that some such relation can be formulated is evidenced by the innumerable “proofs of structure,” which are advanced on the basis of measurements of dissociation constants. However, an examination of the data for different solvents shows that the fundamental assumption that the intrinsic strength of an acid is measured by its dissociation constant in a particular solvent is invalid, since an acid which is stronger than another in one solvent is often weaker in a second solvent; thus in water o -nitrobenzoic acid has a dissociation constant of 6·2 X 10 -3 compared with 1·6 X 10 -3 for 3·5 dinitrobenzoic acid, while in ethyl alcohol the respective constants are 2·42 X 10 -9 and 8·16 X 10 -9 . This fact, which emerges very clearly from the extensive work of Goldschmidt on solutions in methyl and ethyl alcohols and is confirmed by the work of Larsson and of Halford, means that it is impossible to transfer a scale of acidity from one solvent to another, and renders of doubtful significance the rules previously formulated on the basis of results in water