Session DM3: Workshop III: Realistic implementation of plasma‐surface interactions in simulations of technological plasmas

Live

The interactions of plasma particles with the boundary surfaces affect the discharge by particle absorption, emission, reflection, etc. In the simulations, these processes are described by surface interaction coefficients. The surface coefficients are input parameters in the simulations, which are often unknown from measurements, therefore roughly estimated. However, these parameters can largely influence the calculated discharge characteristics. Recently, the importance of the realistic description of the various surface processes in simulations of technological plasmas has attracted increasing attention.\\ The objective of the workshop is to stimulate discussion about the description of various plasma-surface interactions in discharge models, to provide basic theoretical aspects behind the plasma-surface models used in the simulations, theoretical and experimental results on plasma-surface interactions to improve the discharge models, to present numerical challenges encountered in the simulations, experiments with validation efforts and examples of successful validation works.   

Live

Plasma simulations require accurate yield data to predict the electron flux that is emitted when plasma-exposed surfaces are bombarded by energetic particles. One can measure yields directly using particle beams, but it is impractical to create a separate beam of each particle produced by typical plasmas. In contrast, in situ measurements, performed during plasma exposure, provide useful values for effective yield, which includes the effects of all the incident particles. Here, in situ measurements were performed in a radio-frequency (rf) biased, inductively coupled plasma (icp) system in 0.67 Pa and 1.33 Pa (5 mTorr and 10 mTorr) of argon gas. The rf current and voltage across the sheath adjacent to the rf-biased electrode were measured, along with Langmuir probe measurements of ion current density and electron temperature. The measurements are analyzed by a numerical sheath model, which allows the emitted electron flux to be distinguished from other current mechanisms. The effective yield, i.e., the ratio of the emitted electron flux to the incident ion flux is also measured, as a function of incident ion energy. Results for the effective yield of a sputter-deposited SiO2 film are reported and compared with previous work. From the effective yield and additional literature data, recommended emission yields are obtained for each incident particle: photons, Ar$+$ ions, Ar metastables and Ar fast neutrals. Effective emission yields were also measured in mixtures of Ar and CF4. Comparison of these results with mass spectrometer data allows bounds to be placed on the individual emission yields of the most prevalent positive fluorocarbon ions.   

Live

In simulations of technological plasmas, the choice of the heavy-particle induced secondary electron emission (SEE) coefficient ($\gamma$) is crucial. However, there is a lack of data available regarding $\gamma$ for different particle species and surfaces. In order to obtain a realistic value for $\gamma$ in RF plasmas, a computationally assisted spectroscopic technique (the $\gamma$-CAST method) has been introduced [Daksha et al. 2016 \textit{J. Phys. D: Appl. Phys.} \textbf{49} 234001], which is a quantitative implementation of searching for the best match between the spatio-temporal distribution of the excitation rate obtained by phase resolved optical emission spectroscopy (PROES) and the ionization rate obtained by simulations, by varying the SEE coefficient. This method is revisited in this work for CCPs operated in noble gases. The comparison of experimental and computational data results in remarkably different spatio-temporal distributions of the excitation and ionization rates in neon, revealing the limitations of PROES to probe the discharge operation mode. This talk summarizes the applicability of computationally assisted spectroscopic techniques to measure in-situ effective $\gamma$-coefficients for numerical plasma simulations.\\ \\In collaboration with: Aranka Derzsi, Julian Schulze, Ihor Korolov, Peter Hartmann, Zoltan Donko, affiliations with Wagner Research Centre for Physics; West Virginia University, Ruhr-University Bochum, Dalian University of Technology   

Live

In discharge modeling the various plasma surface interaction processes may have a significant influence on the discharge properties. These processes are commonly described by parameters that give the probability of occurrence of the process such as surface recombination to form molecules, surface quenching of metastable states and electron emission from surfaces due to ion, electron and neutral bombardment of surfaces. The surface interaction parameters, often describe a complex processes, that are not well understood, by a single number. These parameters have been measured for some processes, and usually over a limited parameter range, and commonly there are varying values for these parameters that have been measured. Some examples of measured values of such parameters will be discussed, including recombination and quenching probabilities on metal surfaces. The choice of these parameters can have a significant influence on the discharge, including particle densities, power absorption processes and discharge operation. Some examples of the influence of surface quenching of metastables, and surface recombination, on discharge properties and operating mode, and plasma parameters, in capacitively coupled, inductively coupled and magnetron sputtering discharges, will be given.   

Live

The effects of realistic surface coefficients on the charged particle dynamics in geometrically asymmetric capacitively coupled argon discharges operated at low pressure and high voltage are studied by 2D Particle-In-Cell/Monte Carlo collision simulations. By including plasma and SiO$_2$ surface interaction processes, the energy dependent ion ($\gamma$-electron) and electron induced ($\delta$-electron) secondary electron emission are found to influence the electron dynamics a lot. Due to the high energy ion bombardment at the electrodes, a high number of $\gamma$-electrons is emitted. These $\gamma$-electrons are found not to contribute much to the ionization directly, as they are too energetic after being accelerated by the sheath electric field, but they can significantly enhance the $\delta$-electron emission from boundary surfaces, especially in asymmetric discharges, where they constitute the main channel for the $\delta$-electron emission. After the $\delta$-electrons are emitted, they cause nearly 40\,\% of the total ionization. By switching on and off the surface coefficients, large changes of the electron dynamics and a reduced plasma density is found, which indicates that both the $\gamma$- and $\delta$-electrons play important roles in the discharge.\\ \\In collaboration with: Julian Schulze, Ruhr-University Bochum;Peter Hartmann, Wigner Research Centre for Physics; Zoltan Donk\'o, Wigner Research Centre for Physics; Yuan-Hong Song, Dalian University of Technology   

Live

Despite the remarkable progress in ab-initio modelling of plasma-surface interactions, atomistic simulations on time and length scales that are relevant to real experimental systems remain challenging. Therefore, coarse-grained deterministic models are often used in plasma modeling. In deterministic description, surface kinetics is formulated in terms of fractional coverages of different types of active sites and simulated by a system of differential equations. Such mesoscopic approach is simple and computationally efficient but it is inherently approximate because it does not account for the complex microscopic details of the underlying processes. In this presentation, Kinetic Monte Carlo(KMC) modelling of surface kinetics in reactive plasmas is discussed. We demonstrate that KMC is particularly well suited to bridge the gap between the complexity of atomistic simulations and the effectiveness of deterministic models. In KMC approach, the master equation for a given system is not explicitly solved, but instead the underlying Markov process is simulated numerically. KMC algorithms are exact, they can handle reaction probabilities that depend on the local configuration of the system and they are well suited to model heterogeneous surfaces that exhibit distributions of adsorption energy and reactivity. In this report, KMC modelling of atomic recombination and molecular formation on surfaces in contact with reactive plasmas is discussed. Simulations are benchmarked against the available experimental data. In particular, surface processes in systems with a distribution of active sites is analyzed in detail.