A new perspective on aqueous electrolyte solutions

Aqueous electrolyte solutions are central to many natural phenomena and industrial applications. This leads to the continuous development of increasingly complex analytical models to predict their chemical properties. These are all based on an explicit, atomistic description of ion-ion electrostatic interactions combined with mean-field approaches for the dielectric response of water. Such approaches approximate the complex multi-body ion-ion correlations to pair interactions, introducing the concept of ion-pairs. Despite many achievements, these concepts fail to describe situations where ion-ion correlation and specific solvation become relevant, such as for concentrated electrolyte solutions. Here, we propose a change of perspective, by introducing a statistical, coarse-grained view that bypasses the need to define ion pairs, and does not require any prior knowledge of specific solvation. We base our concept on separating the solution into a spherical observation droplet whose size and average composition are fully determined by the solution parameters and the environment of the remaining solution. This allows us to express the droplet-environment interaction in terms of a generalized multipole expansion, i.e. in a convenient, additive way. We applied this approach to 139 electrolytes including some ionic liquids and notoriously complex electrolytes, such as LiCl or ZnCl2. Our model yields a set of analytical functions sharing the same parameters that simultaneously model the activity coefficient, the osmotic coefficient, and water activity, Those parameters give direct access to the radius-dependent partition function around the observation droplet. The functions predict electrolyte behavior over the whole electrolyte mole fraction range, paving the road toward understanding super-saturated and water-in-salt solutions and electrolyte nucleation.