La dispersion turbulente et l'évaporation des particules dans des plasmas à couplage inductif

Abstract

This thesis presents mathematical modeling studies on some basic phenomena in radio frequency (rf) induction plasma and plasma treatment of powder materials. At the outset, a three-equation turbulence fluid model, including plasma density fluctuations, was employed to study turbulence phenomena in an rf plasma discharge. The density fluctuations were found to have negligible effect on plasma temperature and velocity profiles in the present flow conditions, compared to those results obtained using the standard k-[epsilon] turbulence model. It was demonstrated that two distinct regions coexist in the rf plasma torch: the turbulent flow region and the laminar flow region, each having distinct temperature fields and flow patterns. Based on this turbulence study, particle turbulent dispersion in an rf plasma reactor system has been investigated by considering the effect of plasma velocity fluctuations on the particle dynamics. The k-[epsilon] turbulence model was used to describe the plasma phase, while an Eulerian approach was employed to describe the particle dynamics. A Particle-Source-in-Cell (PSI-Cell) model was adopted to represent the plasma-particle interactions. A dimensionless number, the Stokes number, was introduced to elucidate the significance and extent of particle turbulent dispersion. It was shown that particles have similar dispersion properties when the Stokes numbers are similar, regardless of the particle material physical properties and size. Alumina particle heating and evaporation in two types of rf plasma, Ar-H[subscript 2] and Ar-N[subscript 2] , were investigated to show the effects of working gas on the particle evaporation. The predicted particle size distributions for the Ar-H[subscript 2] plasma fitted the measurements better than those for the Ar-N[subscript 2] at higher particle feed rates. Four different expressions for the Nusselt number were used in the modeling to assess their effects on plasma-particle heat transfer models. The modeling results revealed that significant differences exist in the plasma-particle heat transfer rate and in the final particle size distributions using different models. It was also found that a proper plasma-particle heat transfer model can significantly improve the predicted results."--Résumé abrégé par UMI.