Nanofluidics is an exciting interdisciplinary research area that explores the behaviour of fluids under nanoscale confinement. At this scale, molecular effects are expected to dominate the fluid physics, and the concomitant breakdown of the continuum approximation makes the Navier-Stokes equations no longer valid.

Many anomalous transport and thermodynamic phenomena kick in, which are still not fully understood, but that may potentially lead to breakthrough innovations. Examples range from the molecular and Knudsen self-diffusion mechanisms through nanoscale pores, crucial in shale-gas extraction processes, to the velocity slip along solid surfaces, whose better understanding may lead to optimise the design of desalination membranes. In this research, we investigate nanoscale transport phenomena using high-fidelity event-driven molecular dynamics simulations and theoretical analysis, based on the kinetic theory of rarefied and dense fluids.

This work is being conducted by Carlos.

Academic(s) involved: Matthew and Livio

The Boltzmann equation, or any other kinetic model equation, must be supplemented by boundary conditions that model the gas-surface interactions (GSI). These are typically formulated via the so-called scattering kernels and are of paramount importance in modelling non-equilibrium gas dynamics transport. Applications motivated by this research area range from predicting aerodynamic drag forces on satellites operating at very low Earth orbits to enhanced microprocessor cooling and manufacturing, where the gas transport occurs at micro/nanoscale confinements.

This project uses high-fidelity molecular dynamics simulations to accurately resolve the trajectories of molecules in the scattering process and aims to (a) assess the applicability of widely used scattering kernels and (b) propose new scattering models that account for all physical effects that occur at the surface (e.g. physisorption, roughness etc).

This work is being conducted by Yichong.

Academic(s) involved: Matthew and Livio

How gas flows behave in microscale geometries or at low pressure is important to many emerging technologies and engineering challenges. Such examples include: next-generation photolithography machines that manufacture smaller, faster processor chips; thermal management of high-powered electronics using evaporating nanopipe membranes; and high-precision electrospray ionization mass spectrometry. The behaviour of these rarefied flows are, however, not well understood, requiring models of both particle and continuum physics. Existing simulation tools are not fit-for-purpose; they are either too costly to deal with these problems (e.g. DSMC) or inaccurate to reach the full range of Knudsen numbers (e.g. Navier Stokes).

In this research project a new open-source 3D flow model is being developed that uses machine learning techniques to intelligently couple accurate particle solvers to efficient continuum models, obtaining the best-of-both worlds within a single numerical tool.

Researcher(s) involved: Giorgos, Matthew and Livio