Hybrid particle-fluid modeling of plasmas

Summary form only given. Particle in Cell (PIC) methods enjoy great success in modeling devices that include moderately dense plasmas. However, as the plasma density becomes high in a large volume, the number of particles to track becomes computationally prohibitive. Reducing the number of particles by creating larger "macro particles" introduces significant numerical error. Alternatively, high-density plasmas in large volumes can be modeled using a dielectric fluid description. However, such a description is not accurate for non-Maxwellian energy distributions. Some physical processes, such as air breakdown phenomena, involve high-density plasmas with energy distributions that are partially Maxwellian but include a long, high-energy tail. The particles in the high-energy tail are those responsible for the majority of interactions that lead to a qualitative change in the physical behavior. Thus, a fluid description that neglects the high-energy tail fails to capture the physics of interest. A hybrid plasma description has the potential to capture the relevant physics in a tractable computational time. The hybrid description uses a fluid treatment for the particles that fall within the Maxwellian energy distribution, and uses PIC to model the particles in the high-energy tail. The computations are performed using the Improved Concurrent Electromagnetic Particle in Cell (ICEPIC) code. The PIC and fluid models have been tested both independently and in combination. In all cases, the computational techniques were able to accurately reproduce the theoretical curves for the dispersion relation and the transmitted power of electromagnetic plane waves incident on a plasma in a two dimensional box. Adding collisions of the plasma particles with a background gas will affect the transmission of electromagnetic radiation through a dense plasma. We wish to study this effect by adding collisions to ICEPIC; a modification that complicates the modeling because it requires that the- fluid and particle treatments are consistent in the way they interact with the background gas. Consistency will ensure that changes in partial distributions of particles and fluid that make up the hybrid will not result in different predictions for the transmitted power. Future computations that involve a growth of the plasma density over many orders of magnitude will require several adjustments of the particle-fluid balance, so it is important to verify consistency. We test this by investigating the sensitivity of the collisional model to shifts in the particle-fluid balance.