Particle In Cell Simulation Of A Cusp Confined Plasma With Monte Carlo Collisions

The object-oriented paradigm provides an opportunity for advanced PI C modeling, increased flexibility, and extensibility. Particle-in-cell codes for simulating plasmas are traditionally written in structured FORTRAN or C. This has resulted in large legacy codes that are difficult to maintain and extend with new models. In this ongoing research, we apply the object oriented design technique to address these issues. The resulting code architecture, OOPIC (Object-Oriented Particle-in-Cell). is a two-dimensional (x-y, r-z) relativistic electromagnetic/electrostatic PIC-MCC (particle-in-cell, Monte Carlo collisions) plasma simulation. OOPIC includes a growing number of boundary conditions, and can model complicated configurations, including internal structures, without recompilation. it is available to models from DC and RF discharges to high power microwave tubes.

The modeling of collisions among particles in space plasma media poses a challenge for computer simulation. Traditional plasma methods are able to model well the extremes of highly collisional plasmas (MHD and Hall-MHD simulations) and collisionless plasmas (particle-in-cell simulations). However, neither is capable of trealing the intermediate, semi-collisional regime. The authors have invented a new approach to particle simulation called Quiet Monte Carlo Direct Simulation (QMCDS) that can, in principle, treat plasmas with arbitrary and arbitrarily varying collisionality. The QMCDS method will be described, and applications of the QMCDS method as 'proof of principle' to diffusion, hydrodynamics, and radiation transport will be presented. Of particular interest to the space plasma simulation community is the application of QMCDS to kinetic plasma modeling. A method for QMCDS simulation of kinetic plasmas will be outlined, and preliminary results of simulations in the limit of weak pitch-angle scattering will be presented.

Four studies were considered to simulate the ion behavior in the auroral region and the polar wind.

We propose a computational method to self-consistently model small angle collisional effects. This method may be added to standard Particle-In-Cell (PIC) plasma simulations to include collisions, or as an alternative to solving the Fokker-Planck (FP) equation using finite difference methods. The distribution function is represented by a large number of particles. The particle velocities change due to the drag force, and the diffusion in velocity is represented by a random process. This is similar to previous Monte-Carlo methods except we calculate the drag force and diffusion tensor self- consistently. The particles are weighted to a grid in velocity space and associated Poisson equations'' are solved for the Rosenbluth potentials. The motivation is to avoid the very time consuming method of Coulomb scattering pair by pair. First the approximation for small angle Coulomb collisions is discussed. Next, the FP-PIC collision method is outlined. Then we show a test of the particle advance modeling an electron beam scattering off a fixed ion background. 4 refs.

During 20th century few branches of science have proved themselves to be more industrially applicable than Plasma science and processing. Across a vast range of discharge types and regimes, and through industries spanning semiconductor manufacture, surface sterilisation, food packaging and medicinal treatment, industry continues to find new usefulness in this physical phenomenon well into 21st century. To better cater to this diverse motley of industries there is a need for more detailed and accurate understanding of plasma chemistry and kinetics, which drive the plasma processes central to manufacturing. Extensive efforts have been made to characterise plasma discharges numerically and mathematically leading to the development a number of different approaches. In our work we concentrate on the Particle-In-Cell (PIC) - Monte Carlo Collision (MCC) approach to plasma modelling. This method has for a long time been considered computationally prohibitive by its long run times and high computational resource expense. However, with modern advances in computing, particularly in the form of relatively cheap accelerator devices such as GPUs and co-processors, we have developed a massively parallel simulation in 1 and 2 dimensions to take advantage of this large increase in computing power. Furthermore, we have implemented some changes to the traditional PIC-MCC implementation to provide a more generalised simulation, with greater scalability and smooth transition between low and high (atmospheric) pressure discharge regimes. We also present some preliminary physical and computational benchmarks for our PIC-MCC implementation providing a strong case for validation of our results.