Water flow at the nanoscale, a computational molecular and fluid dynamcis investigation

A complete understanding of the most common and abundant fluid on Earth, water, has challenged the scientific community since the dawn of science. Nanoscale confinement and flow through nanotubes show counter-intuitive behaviour and produce interesting new phenomena. Experimental studies of water flow at the nanoscale report diverse results because measurement devices still struggle to provide the necessary accuracy. Molecular Dynamics (MD) simulations can instead provide molecular detail and a plethora of information. The main disadvantage of the method is space and time restraints on the simulated problems. To overcome this, hybrid multiscale methods that combine MD with macroscopic equations have been recently developed. This thesis investigates water flow at the nanoscale, over free surfaces and in nanotube membranes, using MD, multiscale methods and macroscopic hydrodynamic equations. Initially, water nanodroplets on static and moving surfaces of different hydrophilicity are studied here with the developed MD solver. The findings, contrary to macroscopic observations, suggest that the dynamic contact angle of water nanodroplets on graphite is independent of the capillary number. Then, a new method is firstly presented here in order to measure the average molecular orientation of water and highlight any type of anisotropy. For the first time, biaxial ordering of water molecules close to the nanotube walls is observed. It is also found that static electric fields can control the flow rate through them (e.g. a 2 V/nm electric field increases the flow rate by more than 300% for the same pressure difference). The effect of the wall material on water flow is also investigated. The accuracy of a flow enhancement prediction model originally suggested by Mattia et al. [1] is tested, giving satisfactory results for nanotube membranes of small thickness. Finally, results from hybrid multiscale simulations are presented in this thesis, showing perfect agreement with pure MD simulations. The same multiscale method is used for the first time to simulate water flow through a millimeter thick membrane with a pore diameter smaller than 2 nm. All the results presented in this thesis contribute towards the better understanding of water flow at the nanoscale. In parallel this thesis provides a number of computational tools and methods that will help in future studies as well as in the design and simulation of nanoscale fluid devices.A complete understanding of the most common and abundant fluid on Earth, water, has challenged the scientific community since the dawn of science. Nanoscale confinement and flow through nanotubes show counter-intuitive behaviour and produce interesting new phenomena. Experimental studies of water flow at the nanoscale report diverse results because measurement devices still struggle to provide the necessary accuracy. Molecular Dynamics (MD) simulations can instead provide molecular detail and a plethora of information. The main disadvantage of the method is space and time restraints on the simulated problems. To overcome this, hybrid multiscale methods that combine MD with macroscopic equations have been recently developed. This thesis investigates water flow at the nanoscale, over free surfaces and in nanotube membranes, using MD, multiscale methods and macroscopic hydrodynamic equations. Initially, water nanodroplets on static and moving surfaces of different hydrophilicity are studied here with the developed MD solver. The findings, contrary to macroscopic observations, suggest that the dynamic contact angle of water nanodroplets on graphite is independent of the capillary number. Then, a new method is firstly presented here in order to measure the average molecular orientation of water and highlight any type of anisotropy. For the first time, biaxial ordering of water molecules close to the nanotube walls is observed. It is also found that static electric fields can control the flow rate through them (e.g. a 2 V/nm electric field increases the flow rate by more than 300% for the same pressure difference). The effect of the wall material on water flow is also investigated. The accuracy of a flow enhancement prediction model originally suggested by Mattia et al. [1] is tested, giving satisfactory results for nanotube membranes of small thickness. Finally, results from hybrid multiscale simulations are presented in this thesis, showing perfect agreement with pure MD simulations. The same multiscale method is used for the first time to simulate water flow through a millimeter thick membrane with a pore diameter smaller than 2 nm. All the results presented in this thesis contribute towards the better understanding of water flow at the nanoscale. In parallel this thesis provides a number of computational tools and methods that will help in future studies as well as in the design and simulation of nanoscale fluid devices