The electron energy distribution of a low-temperature dusty plasma has been measured via a Langmuir probe. An unexpected broad peak at energy in the 2–4 V range has been observed. This can be theoretically reproduced for a sufficiently large electron emission rate from the nanoparticles dispersed in the plasma. A careful analysis of the nanoparticle energy balance, using measured values of nanoparticle concentration and plasma density, confirms that particles are sufficiently hot under the conditions of this study to rapidly inject electrons into the plasma via field-assisted thermionic emission. This work suggests that the presence of dust affects the plasma ionization balance more deeply than previously thought.

No abstract

In this contribution, we describe the development of a test-bed for the characterisation of non-thermal dusty plasmas via Langmuir probe. This technique, while allowing the precise determination of plasma parameters and electron energy distribution function (EEDF), is notoriously difficult to apply in dust-forming chemistries. We overcome this limitation by utilising a two-plasmas system in which the particle precursor, in this case acetylene, is fully consumed and converted into nanoparticles in a first plasma reactor, followed by the injection of the dust into a second plasma reactor where the Langmuir probe measurement is performed. This approach allows studying the influence of the variation of process parameters on the dusty plasma properties, all while leaving the nucleation and growth phase of the particles unaffected and fully decoupled from the discharge in which the measurement takes place. We have applied this approach to the case of graphitic carbon nanoparticles dispersed in an argon-hydrogen mixture. We have monitored the quality of the Langmuir probe measurement, and found that it is minimally affected by the presence of the graphitic particles even after several EEDF measurements. Our measurements confirm the unipolar charging of nanoparticles in non-thermal plasmas, consistent with previous observations and theoretical predictions. We also observe an unexpected trend with plasma input power: the charge carried by the particles does not monotonically increase with increased power, instead starts decreasing at sufficiently high input power levels.

This paper discusses the use of probabilistic deep neural networks for the prediction of the electron energy probability function in low-temperature non-thermal plasmas. The neural networks are trained using optical emission spectroscopy and Langmuir probe measurements, with the goal of providing a reliable estimate of the electron energy probability function solely from optical emission data. The performance of both non-Bayesian and Bayesian networks is evaluated. It is found that Bayesian models are preferable as they assign a higher level of uncertainty to their prediction especially when the dataset used to train them is small. This work describes one of the many potential applications of machine learning in plasma science and technology.