Assessing the Impact of Cross-Section Data Variability on Particle-in-Cell Monte Carlo Collisions (PIC-MCC) Simulations of Low-Temperature Plasmas (LTP) via Uncertainty Quantification

Accurate simulation of LTPs using PIC-MCC method depends on the fidelity of input collision cross-section (CS) data. For many gases, multiple CS datasets are available from experimental, theoretical, and numerical sources, resulting in significant variability . The lack of universally accepted benchmark data remains a major challenge. In this study, we conduct a systematic uncertainty quantification (UQ) analysis to assess the sensitivity of 2D-3V PIC-MCC simulations of generic ExB plasma based systems to CS dataset variability. We focus on two representative gases, Argon and Hydrogen, and perform simulations under identical conditions across a range of magnetic field profiles using multiple datasets from the LXCat platform [1]. Electron density, potential, temperature profiles, and energy distribution functions (EDF), are analyzed. A bootstrapping-based UQ approach is used to compute 95% confidence intervals, revealing that simulation output variability is strongly dependent on gas type, spatio-temporal conditions, and magnetic field strength.Our simulation results show that even widely used gases can exhibit discrepancies in plasma behavior due to CS data selection and underscore the need for standardized, uncertainty-tagged cross-section datasets for improving the reliability of LTP simulations and future efforts towards validation and refinement of open-access CS databases.

[1] https://lxcat.net/