Extending first principle plasma-surface simulations to experimentally relevant scales. (25th June 2018)

• Record Type: Journal Article
• Authors: Bonitz, M; Filinov, A; Abraham, J W; Loffhagen, D
• Is Part Of: Plasma sources science & technology. Volume 27:Number 6(2018:Dec.)
• Publication Date: 2018-06-25
• Subjects: plasma-surface interaction -- molecular dynamics -- ab initio simulation -- multi-scale modeling; Plasma (Ionized gases) -- Periodicals; 530.44
• DOI: 10.1088/1361-6595/aaca75
• Languages: English
• ISSNs: 0963-0252

Abstract

Abstract: The physical processes at the interface of a low-temperature plasma and a solid are extremely complex. They involve a huge number of elementary processes in the plasma, in the solid as well as charge, momentum and energy transfer across the interface. In the majority of plasma simulations these surface processes are either neglected or treated via phenomenological parameters such as sticking coefficients, sputter rates or secondary electron emission coefficients. However, those parameters are known only in some cases, so such an approach is very inaccurate and does not have predictive capability. Therefore, improvements are highly needed. In this paper we briefly summarize relevant theoretical methods from solid state and surface physics that are able to contribute to an improved simulation of plasma-surface interaction in the near future. Even though the (quantum–mechanical) equations of motion for the participating charged and neutral particles are known, in principle, full ab initio quantum simulations are feasible only for extremely short times and/or small system sizes. A substantial simplification is achieved when electronic quantum effects are not treated explicitly. Then one arrives at much simpler semi-classical molecular dynamics (MD) simulations for the heavy particles that have become the main workhorse in surface science simulations. Using microscopically (i.e., density functional theory) founded potentials and force fields as an input, these MD simulations approach the quality of ab initio simulations, in many cases. However, despite their simplified nature, these simulations require a time step that is of the order or below one femtosecond making it prohibitive to reach experimentally relevant scales of seconds or minutes and system sizes of micrometers. To bridge this gap in length and time scales without compromising the first principles character and predictive power of the simulations, many physical and computational strategies have been put forward in surface science. This paper presents a brief overview on different methods and their underlying physical ideas, and we compare their strengths and weaknesses. Finally, we discuss their potential relevance for future plasma-surface simulations. The first class are 'acceleration' techniques that include metadynamics, hyperdynamics, temperature accelerated dynamics, collective variable driven hyperdynamics and others. Recently we have presented a novel approach: selective process acceleration (Abraham et al 2016 J. Appl. Phys. 119 185301 ) which we discuss in some more detail. The second promising route to longer accurate simulations is dynamical freeze out of dominant modes which we have introduced recently for the simulation of neutral atom sticking on a metal surface (Filinov et al 2018 Plasma Sources Sci. Technol. 27 064003 ). In this article we give a more general view on this method that allows to accurately combine first principles MD simulations with semi-analytical models and discuss possible applications that are of potential relevance for plasma physics.