Computational modeling of low-density ultracold plasmas

Abstract

In this dissertation I describe a number of different computational investigations which I have undertaken during my time at Colorado State University. Perhaps the most significant of my accomplishments was the development of a general molecular dynamic model that simulates a wide variety of physical phenomena in ultracold plasmas (UCPs). This model formed the basis of most of the numerical investigations discussed in this thesis. The model utilized the massively parallel architecture of GPUs to achieve significant computing speed increases (up to 2 orders of magnitude) above traditional single core computing. This increased computing power allowed for each particle in an actual UCP experimental system to be explicitly modeled in simulations. By using this model, I was able to undertake a number of theoretical investigations into ultracold plasma systems. Chief among these was our lab's investigation of electron center-of-mass damping, in which the molecular dynamics model was an essential tool in interpreting the results of the experiment. Originally, it was assumed that this damping would solely be a function of electron-ion collisions. However, the model was able to identify an additional collisionless damping mechanism that was determined to be significant in the first iteration of our experiment. To mitigate this collisionless damping, the model was used to find a new parameter range where this mechanism was negligible. In this new parameter range, the model was an integral part in verifying the achievement of a record low measured UCP electron temperature of 1.57 ± 0.28K and a record high electron strong coupling parameter, Γ, of 0.35 ± 0.08. Additionally, the model, along with experimental measurements, was used to verify the breakdown of the standard weak coupling approximation for Coulomb collisions. The general molecular dynamics model was also used in other contexts. These included the modeling of both the formation process of ultracold plasmas and the thermalization of the electron component of an ultracold plasma. Our modeling of UCP formation is still in its infancy, and there is still much outstanding work. However, we have already discovered a previously unreported electron heating mechanism that arises from an external electric field being applied during UCP formation. Thermalization modeling showed that the ion density distribution plays a role in the thermalization of electrons in ultracold plasma, a consideration not typically included in plasma modeling. A Gaussian ion density distribution was shown to lead to a slightly faster electron thermalization rate than an equivalent uniform ion density distribution as a result of collisionless effects. Three distinct phases of UCP electron thermalization during formation were identified. Finally, the dissertation will describe additional computational investigations that preceded the general molecular dynamics model. These include simulations of ultracold plasma ion expansion driven by non-neutrality, as well as an investigation into electron evaporation. To test the effects of non-neutrality on ion expansion, a numerical model was developed that used the King model of the electron to describe the electron distribution for an arbitrary charge imbalance. The model found that increased non-neutrality of the plasma led to the rapid expansion of ions on the plasma exterior, which in turn led to a sharp ion cliff-like spatial structure. Additionally, this rapid expansion led to additional cooling of the electron component of the plasma. The evaporation modeling was used to test the underlying assumptions of previously developed analytical expression for charged particle evaporation. The model used Monte Carlo techniques to simulate the collisions and the evaporation process. The model found that neither of the underlying assumption of the charged particle evaporation expressions held true for typical ultracold plasma parameters and provides a route for computations in spite of the breakdown of these two typical assumptions.