betaFeedback

WorksPhysics and Astronomy

Learning Similarity-Reduced Ion Kinetics across Scales and Regimes in Capacitively Coupled Plasmas

Abstract

Systematic kinetic investigations of discharge-mode and heating-mechanism transitions in low-pressure capacitively coupled plasmas (CCPs) are computationally expensive because particle-in-cell/Monte Carlo collision (PIC/MCC) simulations must resolve the inherently slow ion relaxation. In this Letter, we propose a similarity-reduced ion-kinetic representation and a machine learning (ML) reconstruction method for weakly ionized Boltzmann-Poisson discharge systems. Using discharge similarity, the three-dimensional control parameter space, comprising gas pressure p, gap distance d, and driving frequency f, is reduced into a two-dimensional similarity space (pd, fd). A dual-branch ML model predicts the ion density profile at the reference scale and the conditional ion velocity distribution. Taking a 1D3V argon discharge as the paradigm, we use the reconstructed ion kinetic state to initialize the PIC/MCC simulation, which then relaxes to a self-consistent periodic steady state after only a brief evolution. Compared to conventional uniform initializations, this approach achieves a maximum computational speedup of 56.8 times and a median speedup of 27.7 times. Furthermore, the physical regimes of complex nonlinear discharge transitions, including the alpha-gamma mode transition, local-to-nonlocal electron collisional heating mode transition and low-density jumps, are accurately mapped within the similarity space. Similarity transformations extend these boundaries across scales, opening a new avenue for rapid kinetic characterization and optimization of radio-frequency plasma sources.

Paper details

vlnc:W_k9prmutx
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
arXiv nonexclusive license
Usage

views · 6

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…