The SOFTWATER project achieved remarkable progress beyond the state of the art in understanding water and water-like systems, introducing innovative theoretical models, computational techniques, and frameworks that have redefined how we study this fundamental substance.
One of the key breakthroughs was the development of the first theoretical model that integrates both static and dynamic anomalies of water: the hierarchical two-state model. This model provides a unified explanation for water’s unique behavior, attributing its thermodynamic anomalies to the temperature- and pressure-dependent population of locally favored structures and its dynamic anomalies to the interplay between these structures.
The project also introduced the concept of water as a “half-empty liquid”, redefining water as a new class of potential. This mean-field model captures the influence of directional bonding and topological constraints imposed by the tetrahedral geometry, accounting for water’s unique anomalies, the liquid-liquid transition, and its crystallization behavior.
A pioneering achievement of the project was the first application of neural networks to study locally favored structures in water and amorphous ices. This computational approach enabled the identification of structural features that are challenging to detect using conventional methods.
Complementing this was the development of a novel neural network potential for water, incorporating advanced atomic fingerprints that account for both two- and three-body interactions. This potential demonstrated exceptional accuracy in reproducing properties beyond its training set, including dynamic behaviors, thermodynamic anomalies, and the stability of crystalline phases. The neural network potential represents a significant step forward in computational modeling, offering unprecedented precision in capturing water’s complex behavior.
Finally, the project advanced self-assembly science by creating the SAT-assembly framework, a versatile tool for solving the inverse self-assembly problem. This framework enabled the precise design of systems to self-assemble into targeted structures, including the successful crystallization of a colloidal pyrochlore lattice.
Collectively, these breakthroughs represent a significant leap beyond the existing state of the art, providing novel theoretical and computational tools that pave the way for future research into water and related systems. The methodologies developed in the SOFTWATER project have far-reaching implications for condensed matter physics, materials science, and the design of complex self-assembling systems.
