Kinetic analysis of the collisional layer

Thesis

In magnetic confinement fusion devices, plasma-wall interaction sets the boundary conditions for the plasma. As charged particles follow the field lines, they come in contact with the wall and charge it. For many angles of interest, between the magnetic field and the wall, the wall charges negatively due to electrons being lighter than ions. An absorbing wall sets a boundary condition for the ion distribution function of no ions coming back from the wall. This has a strong effect on the plasma dynamics in the scrape-off layer. Here Braginskii fluid equations are often used to model the plasma. This assumes the distribution function to be approximately a Maxwellian, making it impossible to impose the wall boundary condition of no ions coming back. In this thesis we analyse the collisional layer. This is a layer that connects the Maxwellian plasma far away from the wall to the magnetic presheath -- a layer that develops near the wall due to the finite size of ion gyro orbits. To analyse the ion dynamics in the collisional layer, we solve the steady state electrostatic ion drift kinetic equation in one spatial dimension using the full Fokker-Planck collision operator, together with quasineutrality and adiabatic electrons. Absorbing wall boundary condition is used at one end of the spatial domain and, to match with Braginskii's equations, the distribution function for incoming velocities far away from the wall is a Maxwellian. We prove that the kinetic Chodura condition is satisfied at the entrance of the magnetic presheath and that the potential scales as a square root of the distance from the wall near the presheath entrance. We also show that, at the entrance of the collisional layer, the flow of ions has to be supersonic. We have written a semi-Lagrangian finite element code to solve our system of equations numerically. We are able to provide the distribution function at the magnetic presheath entrance and the potential drop across the collisional layer. This can be used to find the distribution function at the wall, which is needed to calculate sputtering yields.

Authors: Abazorius, M

• DOI: 10.5287/ora-vmoayxeoo
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English