Dissipative Particle Dynamics Simulation of Suspensions Rheology, and Electroosmotic Flow in Nanochannels

Abstract

The dissipative particle dynamics (DPD) method is developed using innovative numerical techniques and extensively examined in the contexts of rheology and electroosmosis. In Chapters 3-5, it is attempted to classify practical ranges of DPD parameters under a variety of simulation settings, thermostating schemes and shearing methods. Through a calibration process, useful windows of parameters are categorised so that DPD users can model a wide range of rheological systems conveniently with proper temperature control and equilibrium statistics. DPD was found to perform poorly under certain dissipation rates and shear rates when sheared via original Lees-Edwards boundary condition. Hence, a modified version of this shearing method is shown to be an effective remedy to improve the hydrodynamics and thermal stability of sheared DPD systems. These achievements shed light on unclear correlations between input parameters and simulation outputs, and relatively rectifies the lack of predictability embedded in DPD method. In Chapter 6, it is shown that plain DPD is inherently a flexible numerical tool to reproduce experimental behaviour of dilute to dense suspensions. This is achieved via a simple calibration of parameters without unnecessary and computationally intensive modifications to DPD underlying formulas. In Chapter 7, contrary to existing DPD modellings of electroosmotic flow (EOF), soft-core electrostatic interactions are treated fully explicitly by inclusion of charge clouds around DPD soft beads and adopting the corrected Ewald sum method (EW3DC). The developed DPD platform is then calibrated to match the results of molecular dynamics, and reproduce experimental trends. A new system of unit conversion between DPD reduced units and SI units is introduced, which is also useful in other electrokinetic applications. The coarse-graining degree of beads is set to unity to challenge DPD performance in the smallest possible length scale, i.e. in a nanochannel sized at 3.8 nm.