A new method to determine the arbitrary electron energy distribution function (EEDF) from the optical emission spectroscopic measurement in atmospheric-pressure plasma is introduced. The optical emission spectroscopy (OES) continuum emission spectrum, dominated by electron-neutral bremsstrahlung radiation, is analyzed to inspect the usefulness of the conventional OES measurement range for EEDF determination. The EEDF is reconstructed from the OES continuum radiation spectrum by applying machine learning to solve the bremsstrahlung emissivity equation inversely. Through iterative statistical analysis, the presented genetic algorithm can locate the EEDF reliably. Verification of the algorithm shows that theoretical Maxwellian and Druyvesteynian EEDFs can be partially reconstructed from a realistic OES measurement range. Furthermore, preliminary experimental EEDF results of an argon dielectric barrier discharge (DBD) OES measurement are given. The electron energy range and resolution of the determined EEDF are discussed. The results in this paper show potential for accurate determination of the arbitrary EEDF in atmospheric-pressure plasma using simple OES equipment.

The arbitrary electron energy distribution function (EEDF) of a cold atmospheric-pressure plasma is determined from optical emission spectroscopy measurement. Using the electron-neutral bremsstrahlung-dominated continuum spectrum and relevant cross section data, a partial arbitrary EEDF can be reconstructed through the reinforcement learning-based visible bremsstrahlung inversion (VBI) method. From an atmospheric-pressure argon dielectric barrier discharge plasma emission spectrum (440-900 nm), a partial arbitrary EEDF can be determined with a resolution of 0.2 eV. Resemblance between the obtained arbitrary EEPF and a two-temperature Maxwellian was found with T_e = 0.26 and 2.2 eV. The EEDF range is mostly constrained by the spectrum wavelength range.

Partial arbitrary electron energy distribution function (EEDF) results of cold atmospheric-pressure plasma are reported. The EEDF is obtained through the visible bremsstrahlung inversion (VBI) method. This machine learning method requires only optical emission spectroscopy measurement and a momentum transfer cross section to determine a partial EEDF. Numerical EEDF of a pure argon dielectric-barrier discharge dataset with changing peak-to-peak voltage and a helium-argon discharge with changing mixture ratio are reported. Resemblance between the numerical EEDF and a two-temperature Maxwell distribution are observed and a simplified 3-point numerical EEDF is obtained. The electron temperature and relative electron number density for the bulk and high-energy electron populations are measured. The bulk electron temperature was consistently 0.3 eV. For pure argon, the high-energy electron temperature decreased exponentially from 3-2.2 eV with increasing peak-to-peak voltage from 3.6-6.3 kV. For the helium-argon dataset, the high-energy electron temperature decreased linearly from 4.2-2.2 eV with increasing argon fraction 25%-100%. From an OES measurement, the arbitrary EEDF can be observed by utilization of the VBI method. Based on this numerical EEDF, appropriate assumptions can be applied to simplify the quantification of electron diagnostics.

We present a detailed characterization of an atmospheric pressure plasma jet that produces the metastable oxygen states O(1S) and O(1D) with emission of the “auroral” green line. The device used 99.999% pure argon as a working gas for the plasma generation. Optical emission spectroscopy was used to understand the active species present in the plasma jet and to infer the mechanism of O(1S) creation and destruction. The continuum spectrum was used in conjunction with a method based on machine learning for determining the arbitrary electron probability distribution function. Discharge and plasma properties were estimated from Lissajous plots and using calculations with the BOLSIG+ software. The metastable oxygen forms for all operating parameters of the plasma jet system in a linear electrode configuration. The camera images provided information of the overall plasma jet behavior as parameters were altered. While the metastable oxygen was produced for every iteration, the plasma jet behavior changed considerably when the powered and grounded electrodes are switched. Small admixtures of oxygen and nitrogen were introduced in the plasma jet to understand the kinetic processes of metastable oxygen destruction and the 557.7 nm auroral line. This behavior has implications for plasma reactive chemistry in fundamental areas such as auroral physics as well as technological applications of plasmas.