In several fields, novel low-temperature non-equilibrium atmospheric-pressure plasma applications are being developed. In this type of plasma, mainly electron collisions are causing the desired reactions. Therefore, the shape of the electron energy distribution function (EEDF) is essential for understanding and predicting the plasma behavior. The EEDF shape generally does not resemble a Maxwellian or Druyvesteynian. Instead, determination of arbitrary EEDF should be considered. Existing methods to determine the arbitrary EEDF can either not yet be applied to atmospheric-pressure plasma (probe measurement) or are not very accessible (Thomson scattering). In this work, a novel method to determine arbitrary EEDF from optical emission spectroscopy is presented. By using the continuum emission spectrum, dominated by neutral bremsstrahlung, a machine learning scheme can obtain the arbitrary EEDF in the low energy region through reinforcement learning. Further analysis of the machine learning output allows further extraction of electron energy distribution information. The current capabilities of this method will be discussed together with expected limitations and future improvements.
各種検索
サポート
ヘルプ