Exploration of ice growth through molecular dynamics simulations
MetadataShow full item record
AbstractIce is the solid form of water, the most important chemical compound for life. A large number of atmospherically and biologically relevant processes occur at interfaces between these two phases. At the molecular level, crystallization in general, and ice growth in particular, is a less complex example of a natural process of self-assembling, where an ordered crystal is created from a disordered and mobile liquid. This thesis describes efforts to extend our understanding of the process of ice crystal growth by employing the technique of molecular simulations. Molecular simulations have become a de facto standard for these kinds of studies due to fundamental technical difficulties for experimental methods to probe growing interfaces. The study described in this thesis was done as a series of self-contained and relatively independent investigations linked together by one general goal of extending our understanding of the ice growth process. A new general simulation code was developed to answer technical demands of the project. This simulation code was used to perform all the simulations reported in here. The formal development necessary for this work lead to the publication of two new methods of integration of rotational equations of motion, as well as new simulation and data analysis techniques. An investigation of the diffusive behaviour of the TIP4P-2005 model of water was necessary for interpretation of our initial ice growth study and resulted in another research project component; which results provided information necessary for the analysis of ice growth kinetics and also revealed new details of the translational and rotational dynamics of the TIP 4P- 2005 model in liquid phase. The main body of work directly addressing the primary objective of the project, the process of ice growth, was done as four separate simulation studies which are described in this thesis in detail. The main results of this thesis can be summarized as follows. The molecular dynamics simulation studies of ice crystal growth performed in this work significantly improve the general understanding of thermodynamics and kinetics of crystal formation. It is shown that ice growth is a stochastic process biased by the supercooling at the interface between ice and water. The bias due to the temperature can be clearly seen at time scales above 1 ns. At time scales shorter than 1 ns, the interface fluctuations are strongly affected by the specific structure of the interfacial region. It is found that the molecular mobility plays an important role in the crystal growth process, where at lower temperature molecular diffusion becomes the limiting factor for growth. It is shown that ice formation at different crystal faces of hexagonal ice follows different mechanisms: a layer-by-layer mechanism at the basal face, three dimensional collective mechanism at the secondary prism face, and the prism face grows by utilizing either of these two pathways. The foundation of insights into the process of ice growth provided by this work can be used to continue the quest for a full understanding of crystallization and self-ordering of matter.
Bibliography: p. 295-312