Simulations of bulk and confined bubble nucleation

The present thesis investigates, with atomistic simulations, vapor nucleation and liquid dynamics under nanoscale confinement. The main objective of this work is to go beyond the quasi-static classical picture of liquid-vapor phase transition, including kinetic and inertial effects. The performed simulations provide an accurate description of the phenomenon and a framework to interpret experimental observations. The dynamics of vapor nucleation is investigated in the pure bulk liquid and in confined conditions. In the last case, also wetting transition is studied. Particular attention is devoted to surfaces that combine textured geometries with an hydrophobic chemistry. These are able to stabilize vapor phase within surfaces asperities, producing a state in which liquid is suspended above the entrapped vapor pockets. In these conditions, remarkable properties arise that are collectively known as superhydropobicity. In this suspended state, known also as Cassie-Baxter state, the contact area between solid and liquid is reduced with respect to a flat surface and with respect to the textured surface in which the corrugations are flooded with the liquid. Moreover, the liquid presents a higher contact angle (CA), with a lower CA hysteresis and a reduced liquid-solid friction. Due to these properties, superhydrophobic surfaces are suitable for applications such as self-cleaning glass, window, andwallpaint. Theypreventmoistureaccumulation, helpanti-icing, andallowdropwisecondensationtoincreasetheheattransferefficiencyandwaterharvesting. These are all in-air applications. However, the presence of a large shear free liquid/gas interface suggested that super-hydrophobic surfaces can be used in many submerged applications, e.g. drag reduction, anti-friction, anti-adhesive, anti-corrosion, and boiling heat transfer. Cassie-Baxter state can be destabilized by changes in pressure and temperature, that produce the intrusion of the liquid within surface defects. The corresponding state in which the surface is completely wetted is known as Wenzel state. The loss of super-hydrophobic properties (Cassie-Baxter to Wenzel transition) has proved to be experimentally irreversible. It is therefore crucial to characterize both wetting and recovery mechanisms in order understand how to design surfaces supporting a robust Cassie-Baxter state, i.e. a suspended state that can resist to temperature and pressure fluctuations. Wetting transition and recovery of superhydrophobic state take place via vapor/liquid and liquid/vapor phase transitions occurring under confinement at the nanoscale within geometric defects. Over the last decades, a significant amount of experimental and theoretical work has been devoted to the study of confined liquidvapor transition. In spite of this, not much is known yet about the kinetics of the process. The contribution to the topic obtained during the three years of my PhD is presented in this thesis. The first part of the work has been devoted to develop and test Molecular Dynamics and Monte Carlo methods able to properly simulate multiphase systems. Indeed, it has been demonstrated that serious issues arise when the standard global barostats, developed to simulate bulk systems, are straightforwardly applied to systems with subdomains at different pressures, e.g. liquid and vapor domains during nucleation. A solution to overcome these artifacts has been proposed, consisting in the implementation of a local barostat that imposes a local force balance between a piston and the contacting liquid. With this approach, a more accurate prediction of the vapor nucleation barrier in a super-heated liquid has been obtained. Secondly, the simulation techniques developed at the first stage of my PhD work have been employed to study homogeneous bubble nucleation. At the liquid pressure andtemperaturehereinvestigated, thisphenomenonisarareevent: thewaitingtime to observe the inception of vapor formation is order of magnitude longer than the typical time that can be explored by atomistic simulations. This issue, that causes waste of computational resources, has been tackled by carefully selecting special techniques able to preserve kinetic and inertial effects during bubbles growth. With this approach, “dynamical” quantities have been estimated, e.g. the nucleation rate. Other two essential aspects have been addressed: the limits of theoretical expressions routinely used to evaluate the kinetic prefactor in Eyring equation for vapor nucleation; the relation between successful nucleation events and relevant observables, such as temperature and liquid velocity, at beginning and during bubble expansion. The last section of this thesis is focused on heterogeneous nucleation and wetting of super-hydrophobic surfaces. Recent theoretical and experimental studies have produced conflicting results in the characterization of the pathways by which liquid intrudes in pores. The disagreement resides, specifically, in the symmetry properties expected for the advancing meniscus shape. Experiments show a symmetric pathways, in which the liquid penetrates in the surface pores with an essentially flat meniscus, while quasi-static theories predict that the asymmetric pathway is more probable, in which the liquid entering in the surface cavities bend forming a bubble in a corner. My simulations have proved that inertial effects change the wetting and recovery path with respect the predictions of quasi-static approaches. This reconcile theory and experiments: when the transition is barrierless, as expected in experimental conditions in which only nearly spontaneous processes can be addressed, the more complete theory developed here predicts a symmetric wetting as observed in the experiments.