| Models for low density and high density transport in nanochannels have been developed. At low density, the model incorporates the effect of angle of impact on the average axial momentum loss a gas molecule experiences when moving through a nanochannel. Combined with the temperature dependency on oscillation modes, the model predicts axial diffusivities that compare favorably with simulation results. The model was tested over a range of tube geometries and temperatures. The effect of convection on low density molecular motion also was examined. The random disorder of the system dominates much of the results, but the model follows simulated data reasonably well when at moderate and high temperatures and it performs significantly better than competing models.; The low density transport model predicts that at temperatures less than 600 K, argon gas molecules in a carbon nanotube oscillate within a potential well of -5 kJ/mol depth. With increased thermal energy, the molecular diffusivity increases more quickly than predicted by other models. Axial transport occurs in two domains; a near wall domain, in which there is a significant retardation of flow due to wall interaction, and a ballistic domain, in which axial motion is unaffected by the wall potential.; Fluid structure and transport properties have been examined for several density and channel configurations in higher density systems. In higher density fluid systems, radius of curvature plays a crucial role in determining fluid mobility. Fluid in a cylinder is far more stratified than fluid between two plates. Momentum is thus transferred more effectively between a tube wall and its interior fluid as the radius is decreased. A novel, semi-empirical model that bases viscosity on the extent of fluid ordering has been presented and shown to effectively predict the cumulative atomic flow function. The model requires the cross channel density profile to be known, these were determined by simulation. The dimensionless fitting parameter (chi) for the kinematic viscosity model must be determined with at least one simulation in the systems of interest, chi was set at 0.1. After the value of chi is set, the model applies well across a range of tube geometries from 2 nm in radius to 6 nm in radius. |
