Machine learning for modeling, diagnostics, and control of non-equilibrium plasmas

Abstract

Machine learning (ML) is a set of computational tools that can analyze and utilize large amounts of data for many different purposes. Recent breakthroughs in ML and artificial intelligence largely enabled by advances in computing power and parallel computing present cross-disciplinary research opportunities to exploit some of these techniques in the field of non-equilibrium plasma (NEP) studies. This paper presents our perspectives on how ML can potentially transform modeling and simulation, real-time monitoring, and control of NEP.

Plasma polymerization of hexamethyldisiloxane: Revisited

“…Despite those investigations on HMDSO plasmas for decades, there is still a lack of comprehensive understanding about the reaction pathways and film‐forming mechanisms. Although different ways to improve control of plasma processing are currently investigated, involving modeling and machine learning, [ 12,13 ] we aim to revisit the potential of a much more simplified macroscopic approach and its predictive quality. This approach considers both the initial energy transfer by electrons, described by the electron energy and density, as well as the average energy provided per monomer particle in the plasma as key parameters.…”

On a heavy path – determining cold plasma-derived short-lived species chemistry using isotopic labelling

“…36 With regard to the lifetime of most reactive species described for CAPs, it must be acquiesced that only a small fraction can indeed diffuse into a cell or the cell's vicinity, leaving the question of the ultimate mechanism still open. 37,38 With the controllability of plasma treatments in biomedical applications becoming increasingly relevant to improve safety and efficacy, 39 knowledge of the relevant players in plasma-target interaction, their respective trajectories andmost importantlytheir (bio) chemistry is mandatory. It can be concluded that another route of interplay between the plasma-derived species and the biological system is the covalent modication of biomolecules and the subsequent change of their activity or biological value.…”

A review of the gas and liquid phase interactions in low-temperature plasma jets used for biomedical applications

“…To overcome this challenge, feedback from the sensor to the power source will be essential, yet beyond simply varying the applied voltage or frequency it is not clear how the actual composition of RONS in the plasma could be varied to negate against changes in the external environment. Recent efforts in the area of Machine Learning have certainly shown promise in safety-critical applications, where precise control of plasma parameters is a prerequisite [199]; paving the way for intelligent control of RONS generation.…”