Composite Structural Descriptors Reveal Microscopic Structural Heterogeneity in Liquid Water.

Area of Science:

• Condensed matter physics
• Physical chemistry
• Computational materials science

Background:

• Liquids and glasses lack long-range order but possess short- and intermediate-range structural organization influencing macroscopic properties.
• The precise nature and extent of structural organization in water remain a subject of active scientific debate.
• Understanding local structural order is crucial for characterizing disordered systems.

Purpose of the Study:

• To develop a physically transparent framework for constructing composite structural descriptors.
• To identify optimal descriptors that capture distinct forms of local order in liquids.
• To provide a general strategy for analyzing complex disordered systems.

Main Methods:

• Utilized deep potential molecular dynamics with quantum-chemical accuracy.
• Constructed composite structural descriptors from multidimensional joint distributions of local structural metrics.
• Employed information-theoretic analysis to assess descriptor dimensionality.

Main Results:

• Identified an optimal composite descriptor (Σ) integrating first-shell tetrahedral and second-shell translational order.
• The descriptor robustly separates locally ordered and disordered environments in water.
• Structural populations derived from the descriptor align with the two-state model of water.
• Information-theoretic analysis indicated that two dimensions are sufficient for maximal structural discrimination.

Conclusions:

• The developed framework offers a general strategy for identifying minimal, informative descriptors of local order.
• The optimal composite descriptor provides a robust method for characterizing structural heterogeneity in liquids.
• Findings support a two-state description of water's local structure.