Monte Carlo Algorithms for Low-Temperature Plasmas

Abstract

The main aim of this thesis is to develop and employ Monte Carlo (MC) modelling techniques for the investigation and taming of low-temperature plasmas (LTPs). MC methods simulate physical systems by tracking the temporal of evolution of test particles. The stochastic evolution is performed by generating random numbers sampled from distributions that emulate the underlying physics. In Chapter 2, we present a MC method for solving the complex chemical kinetics of heavy species in LTPs. Additionally, novel variance reduction methods are developed to improve the description of minority species without impacting computation time.

In Chapter 3, we present the first version of the LoKI-MC open-source code, which addresses electron kinetics in a gas discharge subjected to a uniform DC electric field. In Chapter 4, we expand the LoKI-MC capabilities to include anisotropic scattering in any collision type. Moreover, we demonstrate that the inclusion of anisotropic scattering in rotational collisions with H2O molecules is fundamental to obtain accurate agreement between modelling and experiment. In Chapter 5, we extend the formulation to configurations involving AC/DC electric and DC magnetic fields.

The code is thoroughly verified, and novel benchmark calculations are produced. Additionally, we analyze the impact of magnetic fields in detail, distinguishing between configurations with DC and AC electric fields. Finally, in Chapter 6, we consider the rigorous time-dependent MC solution as the gold standard to evaluate the accuracy of two common assumptions for solving space- and time- dependent electron kinetics: the local-field approximation (LFA) and the local-energy approximation (LEA).

The study focuses on homogeneous electron kinetics in nanosecond-pulsed discharges. It is observed that the LEA generally provides more accurate results than the LFA. In general, the methods presented in this thesis allowed for a better understanding of LTPs and an assessment of the accuracy of common approximations used in LTP modelling.