Particle-in-cell Monte Carlo Collision Research Articles

The detachment-induced mode in electronegative capacitively coupled radio-frequency plasmas

Abstract Insights into the spatio-temporally resolved electron power absorption dynamics in capacitively coupled radio-frequency plasmas are essential for understanding the fundamentals of their operation and as a basis for knowledge-based plasma process development. Similar to the γ-mode, an ionization maximum is observed at the sheath edge around the time of maximum sheath voltage in electronegative oxygen discharges at a pressure of 300 Pa. Based on Particle-in-Cell/Monte Carlo Collisions (PIC/MCC) simulations, we demonstrate that this maximum is not only caused by secondary electrons emitted at the electrode and collisionally multiplied inside the sheath. In fact, it also occurs in the complete absence of secondary electrons in the simulation, and is caused by the generation of O− ions by electron attachment close to the electrode during the local sheath collapse. These negative ions are accelerated towards the plasma bulk by the sheath electric field during sheath expansion. By electron detachment from these negative ions, electrons are generated inside the sheath and are accelerated towards the plasma bulk by the instantaneous sheath electric field - similarly to secondary electrons. Ionization is also observed in the plasma bulk and caused by electrons generated by detachment and accelerated by the high drift-and ambipolar electric fields. This detachment-induced electron power absorption is found to have significant effects on the discharge in the presence and absence of secondary electron emission. Its fundamentals are understood based on an analysis of the spatiotemporal electron and O− power absorption dynamics as well as the trajectory of selected O− ions close to the electrode.

Modelling of plasma discharge in a filament negative ion source

Most of the currently operating Negative ion Neutral Beam Injectors (N-NBIs) exploit filament powered sources for the generation of beam ions. Being widely used in fusion experiments, the filament arc technology has been thoroughly investigated and optimized over the years, allowing to achieve excellent performances in terms of extracted beam optics. The source geometry, the magnetic field topology and the arc power strongly influence both the plasma discharge and the background gas properties and, consequently, the beam features. In this framework, this contribution describes a numerical investigation of the plasma properties in a filament-powered negative ion source, performed by means of a 2D3V Particle In Cell-Monte Carlo Collisions (PIC-MCC) code. Specifically, we discuss plasma formation by the thermionic electrons emitted by the filaments, investigating their interaction with the background gas. We also study the plasma diffusion through the magnetic filter field and how the latter modifies plasma density, electron temperature and plasma potential along the axial direction when approaching the plasma facing electrode.

A comparative study on the acceleration techniques for solving finite difference discretization poisson’s equation in the PIC/MCC Method

A wide variety of plasma phenomena have been investigated during the past decades using the particle-in-cell/Monte Carlo collisions (PIC/MCC) method. As an important component of the PIC/MCC method, solving Poisson’s equation is crucial for the accuracy and efficiency of calculations. Different acceleration techniques for solving finite difference discretization Poisson’s equation are investigated and compared, including direct method, iterative method, multigrid (MG) method, parallel computing and inherited initial value. The charge density distribution with a known analytical solution is used to validate the algorithm and code. The optimal relaxation factor for the successive over-relaxation (SOR) method in 2D Poisson’s equation with unequal grid node numbers in different dimensions is derived, which is only related to the dimension with the largest grid number. Although there will be a ‘more optimal’ relaxation factor deviated from in some simulation cases, selecting the optimal relaxation factor derived always leads to a not slow solving speed. However, when SOR is used in MG for smoothing, the optimal relaxation factor will shift to 0.5–1.2 from the theoretical optimal value derived with the increase of MG levels. By comparing the convergence order under different relaxation factors and MG levels, the suitable MG level is proposed as log2[min(N x , N y )]−2. Combining the optimal SOR relaxation factor, MG, parallel computing and inherited initial values, the computational cost may decrease by 5 orders of magnitude than that by the simple Gaussian elimination (GE). Based on the optimal acceleration techniques mentioned above, a benchmark simulation case electron cyclotron drift instability (ECDI) in magnetized plasmas was run to further validate the developed PIC/MCC code. The distributions of electric field in the x-direction, electron density and electron temperature are all consistent with the literatures. This paper provides a reference for the acceleration strategy selection for solving Poisson’s equation quickly in plasma simulations.

Effect of desorbed gas on microwave breakdown on vacuum side of dielectric window

The gas desorbed from the dielectric surface has a great influence on the characteristics of microwave breakdown on the vacuum side of the dielectric window. In this paper, the dielectric surface breakdown is described by using the electromagnetic particle-in-cell-Monte Carlo collision (PIC-MCC) model. The process of desorption of gas and its influence on the breakdown characteristics are studied. The simulation results show that, due to the accumulation of desorbed gas, the pressure near the dielectric surface increases in time, and the breakdown mechanism transitions from secondary electron multipactor to collision ionization. More and more electrons generated by collision ionization drift to the dielectric surface, so that the amplitude of self-organized normal electric field increases in time and sometimes points to the dielectric surface. Nevertheless, the number of secondary electrons emitted in each microwave cycle is approximately equal to the number of primary electrons. In the early and middle stages of breakdown, the attenuation of the microwave electric field near the dielectric surface is very small. However, the collision ionization causes a sharp increase in the number density of electrons, and the microwave electric field decays rapidly in the later stage of breakdown. Compared with the electromagnetic PIC-MCC simulation results, the mean energy and number of electrons obtained by the electrostatic PIC-MCC model are overestimated in the later stage of breakdown because it does not take into account the attenuation of microwave electric field. The pressure of the desorbed gas predicted by the electromagnetic PIC-MCC model is close to the measured value, when the number of gas atoms desorbed by an incident electron is taken as 0.4.

Plasma density enhancement in radio-frequency hollow electrode discharge

The plasma density enhancement outside hollow electrodes in capacitively coupled radio-frequency (RF) discharges is investigated by a two-dimensional (2D) particle-in-cell/Monte-Carlo collision (PIC/MCC) model. Results show that plasma exists inside the cavity when the sheath inside the hollow electrode hole is fully collapsed, which is an essential condition for the plasma density enhancement outside hollow electrodes. In addition, the existence of the electron density peak at the orifice is generated via the hollow cathode effect (HCE), which plays an important role in the density enhancement. It is also found that the radial width of bulk plasma at the orifice affects the magnitude of the density enhancement, and narrow radial plasma bulk width at the orifice is not beneficial to obtain high-density plasma outside hollow electrodes. Higher electron density at the orifice, combined with larger radial plasma bulk width at the orifice, causes higher electron density outside hollow electrodes. The results also imply that the HCE strength inside the cavity cannot be determined by the magnitude of the electron density outside hollow electrodes.

The dynamical behavior of the capacitively coupled argon plasma driven by very high frequency

The dynamical behavior of the capacitively coupled Ar plasma driven by four frequencies (54, 80, 94 and 100[Formula: see text]MHz) at fixed pressure are simulated by the particle-in-cell/Monte-Carlo collisions (PIC/MCC) model based on the electromagnetic field. The magnetic field is generated by plasma current itself. The results show that the electron density, charge density, and electron temperature increase with the frequency increase when radio frequency (RF) power is 40[Formula: see text]W. The electron heating rate and argon ion heating rate near the sheath increase with the frequency increase. The high-energy electron density ([Formula: see text] [Formula: see text]eV) varies nonlinearly with the frequency increase. The high-energy electron density is highest at 80[Formula: see text]MHz among the four frequencies. The electron density, charge density, and high-energy electron density increase with RF power (30, 45, 60, and 75[Formula: see text]W) increase at 80[Formula: see text]MHz. The electron temperature decreases with RF power increase at 80[Formula: see text]MHz. The electron heating rate and argon ion heating rate near the sheath have no change with RF power increase at 80[Formula: see text]MHz. The electron energy probability distributions (EEPF) show that the number of high-energy electrons ([Formula: see text] [Formula: see text]eV) is highest at 80[Formula: see text]MHz. The reason may be the electron resonant heating that occurs in the plasma when the electron cyclotron frequency equals source frequency at 80[Formula: see text]MHz. The electron dynamics of capacitively coupled plasma is important for microelectronics etching and membrane deposition.

A neural-network-based model of radio-frequency hollow cathode discharge characterized using particle-in-cell/Monte Carlo collision simulation

An understanding of the plasma dynamics of radio-frequency (RF) hollow cathode discharges (HCDs) at low to moderate pressures is important due to their wide range of applications. A HCD consists of a hollow cylindrical cavity in the RF-powered cathode separated from a grounded electrode by a dielectric. In RF HCDs, RF sheath heating can play a significant role in plasma production in addition to secondary electrons. In this study, a single hollow cathode hole is modeled using the particle-in-cell/Monte Carlo collision (PIC-MCC) technique at low pressure, where kinetic effects are important. Characterization of a single hollow cathode using PIC-MCC simulation is, however, computationally expensive. For improved computational efficiency, a neural network modeling framework has been developed using the temporal variations of applied RF voltages as input and the electrode current as output. A space-filling design for computational experiments is used, where the variables include the RF voltage at the fundamental frequency, RF voltage at the second harmonic, and their phase difference. The predictions of the electrode current using the trained neural network model compare well with the results of the PIC/MCC simulations, but at a significantly lower computational cost. The neural network model predicts the current very well inside the training domain, and reasonably well even outside the training domain considered in this study.

An experimental and computational study on the ignition process of a pulse modulated dual-RF capacitively coupled plasma operated at various low-frequency voltage amplitudes

The ignition process of a pulse modulated capacitively coupled argon discharge driven simultaneously by two radio frequency voltages [12.5 MHz (high frequency) and 2.5 MHz (low frequency, LF)] is investigated by multifold experimental diagnostics and particle in cell / Monte Carlo collision simulations. In particular, (i) the effects of the LF voltage amplitude measured at the end of the pulse-on period, VL,end , on the spatiotemporal distribution of the electron impact excitation rate determined by phase resolved optical emission spectroscopy, and (ii) the electrical parameters acquired by analyzing the measured waveforms of the plasma current and voltage, are studied. Computed electrical parameters and spatio-temporal excitation maps show a good qualitative agreement with the experimental results. Especially, various breakdown mechanisms are found at different VL,end . At low values of VL,end , the ‘RF-avalanche’ mode (volume process) dominates the electron multiplication process. By increasing VL,end , the ionization caused by the volume electrons is suppressed and the electron loss at the electrodes is enhanced, leading to a delayed ignition. At higher values of VL,end , the avalanche ionization is significantly enhanced by ion-induced secondary electron emission at the electrodes, so that the ignition is significantly advanced.

3D simulations of positive streamers in air in a strong external magnetic field

We study how external magnetic fields from 0 to 40 T influence positive streamers in atmospheric pressure air, using 3D PIC-MCC (particle-in-cell, Monte Carlo collision) simulations. When a magnetic field B is applied perpendicular to the background electric field E , the streamers deflect towards the +B and −B directions which results in a branching into two main channels. With a stronger magnetic field the angle between the branches increases, and for the 40 T case the branches grow almost parallel to the magnetic field. Due to the E×B drift of electrons we also observe a streamer deviation in the opposite −E×B direction, where the minus sign appears because positive streamers propagate opposite to the electron drift velocity. The deviation due to this E×B effect is smaller than the deviation parallel to B . In both cases of B perpendicular and parallel to E , the streamer radius decreases with the magnetic field strength. We relate our observations to the effects of electric and magnetic fields on electron transport and reaction coefficients.

Frequency-dependent electron power absorption mode transitions in capacitively coupled argon-oxygen plasmas

Phase Resolved Optical Emission Spectroscopy (PROES) measurements combined with 1d3v Particle-in-Cell/Monte Carlo Collisions simulations are performed to investigate the excitation dynamics in low-pressure capacitively coupled plasmas (CCPs) in argon-oxygen mixtures. The system used for this study is a geometrically symmetric CCP reactor operated in a 70% Ar-30% O2 mixture at 120 Pa, applying a peak-to-peak voltage of 350 V, with a wide range of driving RF frequencies (2 MHz ⩽f⩽ 15 MHz). The measured and calculated spatio-temporal distributions of the electron impact excitation rates from the Ar ground state to the Ar 2p1 level show good qualitative agreement. The distributions show significant frequency dependence, which is generally considered to be predictive of transitions in the dominant discharge operating mode. Three frequency ranges can be distinguished, showing distinctly different excitation characteristics: (i) in the low frequency range ( f⩽ 3 MHz), excitation is strong at the sheaths and weak in the bulk region; (ii) at intermediate frequencies (3.5 MHz ⩽f⩽ 5 MHz), the excitation rate in the bulk region is enhanced and shows striation formation; (iii) above 6 MHz, excitation in the bulk gradually decreases with increasing frequency. Boltzmann term analysis was performed to quantify the frequency-dependent contributions of the Ohmic and ambipolar terms to the electron power absorption.

Numerical characterization of dual-frequency capacitively coupled plasmas modulated by electron beam injection

The modulated approach of electron beam (EB) injection can achieve favorable parameters for capacitive coupled plasmas (CCP). In this work, a one-dimensional particle-in-cell/Monte Carlo collision (PIC/MCC) model is used to simulate the stable dual-frequency CCP with EB injection. First, when the parameters of EB are kept constant at 0.01 A and 30 eV, the results demonstrate significant enhancements in electron density, self-bias voltage, and ion flux. Furthermore, the electron energy probability function (EEPF) appears to have a transition from a typical bi-Maxwellian distribution to a Maxwellian distribution, and the dominant heating mode shifts from the α-mode to the α-γ-mode. Secondly, when the EB current and energy are all changed, the basic parameters of DF-CCP can be achieved by different modulations. Furthermore, we also discuss the transition of the electron heating mode as the current increases from 0.001 A to 1 A and the energy increases from 10 eV to 490 eV. In particular, we conduct a comparative study among different cases of EB injection. According to these results, the modulation capability of EB injection in DF-CCP is thoroughly investigated, which can greatly benefit atom-scale etching in practical applications.

Study of nanosecond pulsed hollow-cathode electron beam generation and transportation based on PIC-MCC simulation

This work employs the PIC-MCC (Particle-in-Cell Monte-Carlo collisions) method to study the nanosecond pulsed electron beam time-varied beam current profiles through the hollow-cathode discharge plasma evolution and electron beam transportation process. The electron beam’s energy and current distribution characteristics at the anode hole are investigated, and the simulation’s grid and step sizes are set as 0.04 µm and 5×10-13s, respectively. Experimental parameters are used in the simulation, and experimental results are also used to verify the simulation accuracy. The simulations investigate the transportation characteristics of the electron beam under different conditions and obtain the effects of voltage, current, and ion density on the distribution characteristics of the electron beam. The electron beam current generated by the hollow-cathode discharge is gradually transformed from a random distribution in the central region to a Gaussian distribution. Due to the ion-forced channel formed by the runaway electron pre-ionization, the high-energy electron beam achieves focused transport over a certain distance, and the high-energy electron beam has a smaller Gaussian radius than the source input. Besides, the ion-focusing channel formed by the high-energy electron transportation process can realize the transportation of the high-current electron beam in the conduction stage.

Hollow cathode effect in radio frequency hollow electrode discharge in argon

Radio frequency capacitively coupled plasma source (RF-CCP) with a hollow electrode can increase the electron density through the hollow cathode effect (HCE), which offers a method to modify the spatial profiles of the plasma density. In this work, the variations of the HCE in one RF period are investigated by using a two-dimensional particle-in-cell/Monte-Carlo collision (PIC/MCC) model. The results show that the sheath electric field, the sheath potential drop, the sheath thickness, the radial plasma bulk width, the electron energy distribution function (EEDF), and the average electron energy in the cavity vary in one RF period. During the hollow electrode sheath’s expansion phase, the secondary electron heating and sheath oscillation heating in the cavity are gradually enhanced, and the frequency of the electron pendular motion in the cavity gradually increases, hence the HCE is gradually enhanced. However, during the hollow electrode sheath’s collapse phase, the secondary electron heating is gradually attenuated. In addition, when interacting with the gradually collapsed hollow electrode sheaths, high-energy plasma bulk electrons in the cavity will lose some energy. Furthermore, the frequency of the electron pendular motion in the cavity gradually decreases. Therefore, during the hollow electrode sheath’s collapse phase, the HCE is gradually attenuated.

First-principles simulation of optical emission spectra for low-pressure argon plasmas and its experimental validation

Various spectral line emissions are often used for the experimental characterization of low-temperature plasmas. For a better understanding of the relation between the plasma characteristics and optical emission spectra, first-principle numerical simulations for low-pressure radio-frequency driven capacitively-coupled plasmas (CCPs) of argon have been performed by coupling one-dimensional particle-in-cell/Monte Carlo collision (PIC/MCC) simulations with a global collisional-radiative model (CRM). The only ionization and excitation mechanisms included in the PIC/MCC simulations of this study are the electron-impact ionization and excitations of the ground-state Ar atoms, as done commonly, whereas the electron-impact ionization of metastable states and other ionization mechanisms are also included in the CRM to account for the optical emission spectra. The PIC/MCC coupled CRM provides the emission spectra, which are then compared with experimental data obtained from the corresponding Ar CCPs with a gas pressure ranging from 2 Pa to 100 Pa. The comparison has shown good agreement for pressures up to about 20 Pa but increasingly notable deviations at higher pressures. The deviation is ascribed to the missing consistency between the PIC/MCC simulations and CRM at higher pressures, where the ionization from the metastable states is more dominant than that from the ground states, indicating a significant change in the electron energy distribution function due to the electron collisions with excited Ar atoms at higher pressures.

An experimental and numerical study: to investigate the switching characteristics of a field-distortion sparkgap switch in various SF6 admixtures

Pressurized sparkgap switch is one of the major components that is most commonly used in various pulsed-power systems and the stochastic closure time delay (t d ) and jitter (σ d ) in this gas discharge switch has substantial effect on the temporal characteristics of output pulse. Experimental investigations performed with various volumetric concentrations of SF6 in SF6-N2, SF6-dry air and SF6-Ar admixtures evidenced that the intrinsic parameters of the switch such as time delay and jitter were found to be strongly dependent on the nature of gas, volumetric concentration of SF6 in admixture and rate of rise of voltage (dV/dT) of applied trigger pulse for e.g. in SF6/N2 admixture experiments as the concentration of SF6 was increased from 5% to 50% the average time delay and jitter was consequently increased from 278 ns to 413 ns and 17 ns to 56 ns, respectively. The similar increasing trend was observed even in SF6/dry air and SF6/Ar admixture experiments. It was also observed in the investigation that the variation in time delay and jitter is relatively much lesser for the condition when fast rising trigger pulse having slew rates ∼500 V ns−1 is applied than that of trigger pulse with lower dV/dT of ∼3 V ns−1. To further corroborate the stochasticity of streamer discharge in experimented SF6 admixtures, numerical studies have also been carried out by using a Particle-In-Cell/Monte Carlo Collision (PIC/MCC) model. Our experimental and simulation results infer that SF6/Ar admixture with low SF6 concentration is a superior alternative than pure SF6 for pulsed power switching applications.

Breakdown and quasi-DC phase of a nanosecond discharge: Comparison of optical emission spectroscopy measurements with numerical simulations

A nanosecond atmospheric pressure plasma jet operated in pure nitrogen is studied by spatially and temporally resolved optical emission spectroscopy complementing the companion paper (Kuhfeld et al 2023 Plasma Sources Sci. Technol. 32 084001), where the discharge is investigated by means of Particle-in-Cell/Monte Carlo collisions (PIC/MCC) simulations and fluid models. Two temporal phases of the evolution of the discharge are identified: a fast breakdown and a quasi-DC phase. It is shown that during the breakdown phase several ionization waves develop, while after the breakdown the discharge has a structure similar to DC glow discharges, in agreement with the modeling predictions. The results of the measurements of the spatial-temporal dynamics of the light emission are compared with the distribution of densities of the and states reconstructed from the PIC/MCC simulations. A good agreement is demonstrated.

A study of grid failure mode drivers and methods for accelerated life testing of a 30 cm diameter ion thruster

In view of the high cost caused by the 1:1 lifetime verification test of ion thrusters, the lifetime acceleration test should be considered. This work uses the PIC-MCC (Particle-in-Cell Monte-Carlo Collision) method to analyze the five failure factors that lead to the failure of the accelerator grid of a 30 cm diameter ion thruster under the working mode of 5 kW. Meanwhile, the acceleration stress levels corresponding to different failure factors are obtained. The results show that background pressure has the highest stress level on the grid’s erosion. The accelerator grid aperture’s mass sputtering rate under the rated vacuum degree (1 × 10−4 Pa) of 5 kW work mode is 8.78 times that of the baseline vacuum degree (1 × 10−6 Pa), and the mass sputtering rate under worse vacuum degree (5 × 10−3 Pa) is 5.08 times that of 1 × 10−4 Pa. Under the influence of the other four failure factors, namely, the voltage of the accelerator grid, upstream plasma density, the screen grid voltage and mass utilization efficiency, the mass sputtering rates of the accelerator grid hole are 2.32, 2.67, 1.98 and 2.51 times those of the accelerator grid hole under baseline condition, respectively. The ion sputtering results of two 30 cm diameter ion thrusters (both installed with new grids assembly) after working for 1000 h show that the mass sputtering rate of the accelerator grid hole under vacuum conditions of 5 × 10−3 Pa is 4.54 times that under the condition of 1 × 10−4 Pa, and the comparison error between simulation results and test results of acceleration stress is about 10%. In the subsequent ion thruster lifetime verification, the working vacuum degree can be adjusted according to the acceleration stress level of background pressure, so as to shorten the test time and reduce the test cost.

Inertial and anisotropic pressure effects on cross-field electron transport in low-temperature magnetized plasmas

In this paper, a one-dimensional (1D) particle-in-cell Monte Carlo collision (PIC-MCC) model is developed to investigate the effects of anisotropic pressure and inertial terms due to non-Maxwellian velocity distribution functions on cross-field electron transport. The conservation of momentum is evaluated by taking the moments of the first-principles gas-kinetic equation. A steady-state discharge is obtained without any low-frequency ionization oscillations by considering an anomalous electron scattering profile. The results obtained from the 1D PIC-MCC model are compared with fluid models, including the quasi-neutral drift-diffusion (DD), non-neutral DD, and full fluid moment models. The discharge current obtained from the PIC-MCC model is in good agreement with the fluid models. The cross-field electron transport due to the inertial terms, i.e. the gradient of axial and azimuthal drift, is evaluated. Moreover, PIC-MCC simulation results show non-zero, anisotropic, off-diagonal pressure tensor terms due to asymmetric non-Maxwellian electron velocity distribution function, potentially contributing to cross-field electron transport.

On the importance of excited state species in low pressure capacitively coupled plasma argon discharges

In the past three decades, first principles-based fully kinetic particle-in-cell Monte Carlo collision (PIC/MCC) simulations have been proven to be an important tool for the understanding of the physics of low pressure capacitive discharges. However, there is a long-standing issue that the plasma density determined by PIC/MCC simulations shows quantitative deviations from experimental measurements, even in argon discharges, indicating that certain physics may be missing in previous modeling of the low pressure radio frequency (rf) driven capacitive discharges. In this work, we report that the energetic electron-induced secondary electron emission (SEE) and excited state atoms play an important role in low pressure rf capacitive argon plasma discharges. The ion-induced secondary electrons are accelerated by the high sheath field to strike the opposite electrode and produce a considerable number of secondary electrons that lead to additional ionizing impacts and further increase of the plasma density. Importantly, the presence of excited state species even further enhances the plasma density via excited state neutral and resonant state photon-induced SEE on the electrode surface. The PIC/MCC simulation results show good agreement with the recent experimental measurements in the low pressure range (1–10 Pa) that is commonly used for etching in the semiconductor industry. At the highest pressure (20 Pa) and driving voltage amplitudes 250 and 350 V explored here, the plasma densities from PIC/MCC simulations considering excited state neutrals and resonant photon-induced SEE are quantitatively higher than observed in the experiments, requiring further investigation on high pressure discharges.

Numerical Study on Kinetic Effects of Driving Frequency in Atmospheric Radio Frequency Discharges Using Deep Neural Network

The particle-in-cell Monte Carlo collision (PIC-MCC) model is an essential way to investigate the kinetic behaviors of low-temperature plasmas, but usually, it is hugely time consuming in simulating atmospheric radio frequency (RF) plasmas. In this study, a deep neural network (DNN) with multiple hidden layers is developed to predict the kinetic discharge characteristics of atmospheric RF plasmas. The results obtained from the PIC-MCC model are used as the training dataset for the DNN, and the well-trained DNN is able to efficiently yield various kinetic behaviors of atmospheric RF discharges with very high precision. The validation of the results predicted by the DNN algorithm is performed by comparing them with the simulation results directly from the PIC-MCC model. Compared with the time-consuming PIC-MCC simulations, the well-trained DNN takes only 0.01 s to yield the essential kinetic characteristics of atmospheric RF discharges, which saves seven orders of magnitude of computation time compared with the traditional PIC-MCC simulation. The predicted data show that the discharge current of atmospheric RF discharges increases monotonically with the driving frequency at a given applied voltage, and as the driving frequency increases, the electric field in the sheath region is strongly enhanced with the sheath region shrinks and the bulk plasma region expanding. Additionally, the electron energy distribution function (EEDF) can be accurately predicted by DNN; moreover, as the driving frequency increases, the low-energy electrons can be transformed into medium-energy electrons, leading to a transition of the EEDF from a three-temperature distribution to a Maxwellian distribution from the DNN prediction. The study indicates that the well-trained DNN is a promising tool for plasma simulation with very high efficiency and accuracy compared with the PIC-MCC model with huge computational costs, which could provide enough kinetic results to further understand the discharge behaviors in atmospheric RF discharges.