Nanoscale dynamics of ice growth on surfaces

Abstract

The freezing of liquid water to form ice is the most common phase transition on our planet, and the ubiquity of ice growth on surfaces offers many engineering and scientific challenges. Ice growth can prevent the operation of — and cause damage to — a broad spectrum of man-made structures and devices, such as aircraft, ships, wind turbines, photovoltaic devices, heat exchangers, and telecommunications equipment. Consequently, ice growth on surfaces is an important topic of study, and undesirable ice growth can impact both life and property. While there is a vast experimental and numerical literature on how modifying surface characteristics affects ice growth, little is understood about the underlying nanoscale mechanisms. This is because icing is challenging to study experimentally, as the nucleation of ice crystals within super-cooled liquid occur on the order of nanometres. Numerical investigations into icing rely on the use of molecular dynamics (MD) simulations, as MD can accurately resolve the nanoscale molecular interactions relevant to ice nucleation and growth. However, there are two issues with the use of MD: a) these simulations are computationally expensive; and b) as nucleation is a rare event when compared to MD timescales, the simulations need to be accelerated to force ice formation to occur, which affects the accuracy of the results obtained. An alternative seeded MD simulation approach is presented in the present work, which reduces the computational cost while still ensuring accurate simulations of ice growth on surfaces. In addition, this approach enables the study of ice growth on vastly more complex surfaces than have been considered thus far. In this thesis, this approach is used to investigate fundamental questions of surface icing, such as: a) the effect of surface wettability and structure on ice growth; and b) the role of the nanometre-thick interfacial region adjacent to the surface on ice growth, which has been shown to be important in previous experiments. The findings presented here should provide an improved understanding on the role of the surface properties on the structure and dynamics of ice growth, and provide a useful framework for future slab-seeded ice growth simulations.