Homogeneous nucleation of phase transformations in supercooled water

TL;DR

This paper develops a thermodynamic model for phase transformations in supercooled water, explaining the fragile-to-strong transition and predicting properties consistent with experimental data across pressures.

Contribution

It introduces a new enthalpy-based nucleation model that accurately predicts water's phase behavior and transitions, including the fragile-to-strong transition, from known thermodynamic parameters.

Findings

Predicts the fragile-to-strong transition at TLL = 228.5 K.

Describes the formation of a glacial phase at TLL = 0.8367*Tm.

Explains the first-order transition without latent heat at TK2 = 122.4 K.

Abstract

The classical nucleation equation, applied to two liquids, is completed by an additional enthalpy for solid supercluster formation governing the liquid and glass transformations. This model, successfully applied to d-mannitol, triphenyl phosphite and n-butanol, defines a formation rule of strong glacial phase, explaining the origin of the first-order transition of water from fragile-to-strong liquid at TLL = 228.5 K, only knowing Tg = 137.1 K, the melting heat and the melting temperature Tm. All thermodynamic properties and transitions, even under pressure P, are now predicted in agreement with experiments of Kanno and Angell (1979), Mishima (1994), Mishima and Stanley (1998), Loerting et al (2006), Amann-Winkel et al (2013), Shephard and Salzmann (2016, 2017), Tulk et al (2019). This glacial phase is formed at TLL = 0.8367*Tm for P < 0.017 GPa. (TLL) decreases with P < 0.017 GPa and disappears for P > 0.55 GPa. The lowest-density liquid is, at once, the glacial phase of fragile and high-density liquids. It is formed at TLL, remains liquid during the first cooling, and gives rise to the glass phase by heating through a first-order transition without latent heat at TK2 = 122.4 K. Ordered liquid Phase 3 appears during the first heating above Tg = 137.1 K, with its own Kauzmann temperature, superheating above Tm up to Tn+ if crystallization is avoided. The latent heat at TLL exists during the first cooling and is compensated during the next cooling by Phase 3 formation heat. The first-order transition without latent heat at TK2 = 122.4 K induces ordered Phase 3, becomes underlying below Tg in further heating and induces an enthalpy excess at Tm which is used to predict TK2. This description agrees with the broken bond and solid fractal structure percolation theories.