Developing an optimization algorithm for diagnostic modeling of optical emission spectroscopic measurement of non-equilibrium plasmas based on the argon collisional-radiative model

Abstract

In this work, an optimization algorithm was proposed for plasma diagnostic modeling based on a statistical analysis of reduced population density distribution. The algorithm generates a diagnostic equation, whose input parameters are the radiant flux of the multi-optical emission lines, and output parameters are electron temperature T e, electron density N e, and electron energy distribution function (EEDF), based on the dependence of reduced population density onT e, N e, and EEDF. The argon collisional-radiative model and the generalized EEDF were applied for the analysis of the excitation-kinetics. In this study, the diagnostic error for the low-pressure inductively coupled plasma, the low-pressure microwave discharge surface wave plasma, and the atmospheric dielectric barrier discharge plasma were simulated. The simulated diagnostic errors were smaller than those of the previously reported model.

Optical Emission Spectroscopic Measurement of Argon Low-Pressure Inductively Coupled Plasma Based on the Optimization Algorithm of Plasma Diagnostic Model

“…The first goal of this study is to discuss the algorithm for plasma diagnostic modeling. In the previous study [11], we found that the nonmonotony of the reduced population density n i (T e , x, N e )/g i for (T e , x, N e ) caused an increase in the diagnostic error. The algorithm of the previous study did not consider the unimodality of n i (T e , x, N e )/g i on the selection of optical emission lines for diagnosis.…”

“…In the previous study [11], we reported the limitations of the proposed optimization algorithm for a diagnostic model. It was identified that the nonmonotonic dependence of reduced population density n i /g i on unknown parameters (T e , x, N e ) caused local optimal solution and errors of diagnostic results.…”

“…To solve this problem, we proposed an optimization algorithm for the plasma diagnostic modeling of argon (Ar) plasma [11], where the diagnostic model is defined as a fitting equation (objective function) for obtaining the unknown parameters T e , N e , and electron energy distribution function (EEDF) for user-specified plasma conditions (e.g., T e , N e , EEDF, pressure, atomic temperature).…”

“…The aim of the previous reported algorithm [11] is to obtain a more accurate diagnostic model of T e , N e , and EEDF suitable for user-specified plasma conditions (e.g., T e , N e , EEDF, pressure, atomic temperature, and plasma column radius). The diagnostic principle is based on a fitting comparison of reduced population densities between the experimental values by multi-OES line intensities' measurement and calculated values by the collisional-radiative (CR) model.…”

“…An overview of the algorithm flow is as follows: the algorithm calculates the excitation kinetics based on the Ar CR model [12]- [16] and the generalized EEDF [17] (with parameter x (for details, refer to Appendix A); the dependence of reduced population densities on the unknown plasma parameters: n i (T e , x, N e )/g i is analyzed in the algorithm, where i is the level number, n i is the number density of level i , and g i is the statistical weight of level i ; the excited levels which have a large dependency on (T e , x, N e ) are selected: The candidates of fitting equation are generated by combining the selected levels: The simulated diagnostic error of the fitting equation is calculated: The fitting equation with the minimum diagnostic error is selected based on analysis of selfconsistency. For details of the previously reported algorithm, refer to the previous article [11].…”