Fig. 5 | Nature Communications

Fig. 5

From: Machine learning coarse grained models for water

Fig. 5

Post-nucleation ice grain growth simulations performed using ML-BOPdih. System dynamics and evolution of structural motifs of post-nucleation phase after the slow heating from 210 to 260 K. a Snapshots from simulations showing the grain growth process of nanosized grains at 260 K evolving into a single grain. (see Supplementary Movie 2 for a zoom-in view) b Snapshots from simulations (Supplementary Movie 3) showing the time evolution of hexagonal layers in stacking disordered ice. Cubic type molecules are not shown for clarity. Note that the ML-BOP model considers only nearest neighbor interactions but is able to reproduce the random stacking-disorder which is consistent with experimental observations31. c Snapshots from simulations (Supplementary Movie 4) showing the in-plane transformation between cubic and hexagonal layers in stacking disordered ice, viewed along the direction perpendicular to the basal plane of the largest stacking ice grain in the system

Back to article page