SMBP Group Meeting: Ryan Szukalo (Princeton)

Tuesday Apr 15, 2025, 11:00 AM → 12:00 PM America/New_York

3rd Floor Classroom (162 Fifth Avenue )

Speaker: Ryan Szukalo (Princeton)

Topic: Computational Investigation of Water Glasses using Machine Learning Potentials

Abstract: Water's anomalous properties have attracted the attention of scientists for decades, with one explanation being the existence of two metastable liquid states at low temperatures and super-atmospheric pressure, separated by a first-order phase transition. Experimental verification is challenging due to rapid ice formation, with glassy phases acquiring major relevance in experimental, theoretical, and computational explorations of cold, non-crystalline forms of water. We combine machine learning potentials with extensive molecular dynamics simulations to conduct the first investigation of water's glassy states based on quantum mechanical calculations. The models accurately reproduce experimental observations of low- and high-density amorphous ices without being explicitly trained on these states. We also introduce a novel classification technique based on unsupervised dimensionality reduction which allows for microscopic investigation of glassy environments, facilitating analysis of the transition between low- and high-density amorphous ices. Our findings support the liquid-liquid phase transition hypothesis and identify experimental pathways for probing criticality by quenching liquid water near the proposed critical pressure, where enhanced density fluctuations may persist in the glassy state.