Abstract
The mechanism of how ice crystal form has been extensively studied by many researchers but remains an open question. Molecular dynamics (MD) simulations are a useful tool for investigating the molecular-scale mechanism of crystal formation. However, the timescale of phenomena that can be analyzed by MD simulations is typically restricted to microseconds or less, which is far too short to explore ice crystal formation that occurs in real systems. In this study, a metadynamics (MTD) method was adopted to overcome this timescale limitation of MD simulations. An MD simulation combined with the MTD method, in which two discrete oxygen–oxygen radial distribution functions represented by Gaussian window functions were used as collective variables, successfully reproduced the formation of several different ice crystals when the Gaussian window functions were set at appropriate oxygen–oxygen distances: cubic ice, stacking disordered ice consisting of cubic ice and hexagonal ice, high-pressure ice VII, layered ice with an ice VII structure, and layered ice with an unknown structure. The free-energy landscape generated by the MTD method suggests that the formation of each ice crystal occurred via high-density water with a similar structure to the formed ice crystal. The present method can be used not only to study the mechanism of crystal formation but also to search for new crystals in real systems.
Highlights
The mechanism of how ice crystal form has been extensively studied by many researchers but remains an open question
Two discrete radial distribution functions (RDFs) implemented in PLUMED 1.340 were selected as collective variables (CVs), with each represented by a Gaussian window function, w(r), as a function of the oxygen–oxygen distance, r, so that CVs were differentiable with r40
Previous studies indicate that the use of U as a CV is convenient for reproducing crystallization in MTD simulations[26,27]
Summary
The mechanism of how ice crystal form has been extensively studied by many researchers but remains an open question. The timescale of phenomena that can be analyzed by MD simulations is typically restricted to microseconds or less, which is far too short to explore ice crystal formation that occurs in real systems. One of the reasons for this is that appropriate methods for determining the formation mechanisms of ice polymorphs are lacking To resolve this issue, it is necessary to analyze the molecular-scale processes of the structural changes that occur during the formation of ice polymorphs. It is necessary to analyze the molecular-scale processes of the structural changes that occur during the formation of ice polymorphs Computer simulations, such as molecular dynamics (MD), are often useful for this purpose[5]. The timescale of MD simulations is far too short to study the molecular-scale processes of the structural changes that occur in the formation of ice polymorphs.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
