Title:The activity coefficient of $z$:1 ionic solutions scales with the cube root of salt concentration

Abstract:The activity coefficient of ionic solutions was proposed by Bjerrum (1916) and Ghosh (1918) to depend on the cube root of concentration, because of a good fit with data for low and moderate salt concentration. However, the Debye-Hückel (DH) theory that was developed later prevailed, and this theory has a square root dependence. The derivation of the DH equation is not easy or intuitive, and has various uncertain elements, and the fit to data is not very good for salts with non-unity valencies. We develop a model for the activity coefficient based on the Coulomb force between an ion and its nearest-neighbour ions of opposite charge, including the distribution of separations between the ions. For a symmetric salt, we only have to analyze the interactions between one cation and one anion, and we can derive an expression in the dilute limit that depends on the Bjerrum length and the cube root of salt concentration, similar to a result already put forward by Bjerrum. For higher salt concentrations or a non-zero size, numerical evaluation is required. For a 1:1 salt, this cube root dependence describes data of the activity coefficient very well, both at low and intermediate salt concentrations, with also the prefactor correctly predicted. We also analyze an asymmetric 2:1 salt and 3:1 salt which requires numerical solution of all possible orbits of two or three monovalent ions around a central multivalent ion of opposite charge, and this calculation also leads to a very good prediction of the activity coefficient that linearly depends on the cube root of salt concentration up to moderate concentrations. For 2:2 and 3:3 salts, numerical theory again fits data very well, but the cube root limiting law does not apply. The theory can be extended to include ion volume effects, for instance via the Carnahan-Starling equation of state.