Particle‐in‐cell simulation of electron and ion energy distributions in dc/rf hybrid capacitively‐coupled plasmas

Abstract

A Particle‐in‐Cell simulation with Monte Carlo collisions was used to study electron and ion energy distributions (IEDs) in low‐pressure (2.67 Pa) direct‐current (dc)/radio‐frequency (rf) hybrid capacitively‐coupled Ar plasmas. One electrode (dc/rf electrode) of the parallel plate diode was powered by a 13.56 MHz source, and a negative dc bias voltage, whereas the opposite (substrate) electrode was grounded. Secondary electrons emitted from the dc/rf electrode accelerated in the adjacent sheath and entered the plasma, yielding a high‐energy tail of the electron energy distribution. For given dc bias voltage, the plasma density increased as the secondary electron emission yield due to ion bombardment increased. A fraction of the secondary electrons were energetic enough to overcome the sheath potential barrier on the substrate electrode and bombard the substrate. The electron angular distribution on the substrate electrode had a peak of directional electrons superimposed on a typical cosine distribution. The mean energy and angular spread of directional electrons could be controlled by varying the dc bias voltage. However, as the dc bias became more negative, the dc/rf sheath expanded at the expense of the bulk plasma, reducing the plasma density, in agreement with published data. The IED on the substrate electrode exhibited a dominant bimodal feature with multiple shoulder peaks due to ion‐neutral charge exchange collisions. The average ion energy decreased as the dc voltage became more negative, also in agreement with data. Pulsing the plasma power enhanced the tail of the electron energy distribution in the early activeglow (power ON), and yielded a distinct ballistic electron flux on the substrate with energy equal to the applied dc bias. © 2013 American Institute of Chemical Engineers AIChE J, 59: 3214–3222, 2013

Pulsed plasma etching for semiconductor manufacturing

“…Application of power at multiple frequencies (for example dual-frequency [93,107] or even triple frequency capacitively coupled plasmas) and the possibility of pulsing the power at each of these frequencies, introduces a herculean problem in terms of process design and optimization. Another possibility is to use dc pulsing or RF/dc hybrids to better tailor the IED and the EEDF simultaneously [104,108]. The problem of the plethora of knobs is exacerbated by the need to match the different pulsed powers to the plasma.…”

“…PIC-MCC simulations of the IEDFs were in excellent agreement with the experimental data. Diomede et al [103,104] also conducted PIC-MCC simulations of the application of tailored dc voltage steps on an electrode. For the above methodology to work the auxiliary electrode should remain free of any insulating layers that may deposit on the electrode.…”

Cause analysis of the faults in HARC etching processes by using the PI‐VM model for OLED display manufacturing

“…Because of the requirements on high verticality of the ion flux reached on the substrate surface in the HARC etching process, lower frequency bias power is strongly applied to accelerate ions additionally on to the higher frequency main source power applied for the generation of the plasmas, in general. From this nature of the HARC etching process devices, HARC plasmas have the bi‐Maxwellian EEDFs, usually . Measured b were distributed in the very narrow range of 0.715 to 0.775.…”

“…From this nature of the HARC etching process devices, HARC plasmas have the bi-Maxwellian EEDFs, usually. [14,15] Measured b were distributed in the very narrow range of 0.715 to 0.775. This is a very small change according to the value of b itself, but the 0.05 change of shape factor corresponded to a tens of % change in the high-energy tail (region over the effective threshold energy of dissociation or ionization of the process gases, >10 eV in general) for the fixed T e,eff .…”

Argon metastable and resonant level densities in Ar and Ar/Cl2 discharges used for the processing of bulk niobium

“…[4][5][6] The effect of dc biasing on rf CCP etching reactors has been studied by a number of research groups, both experimentally and with particle-in-cell simulations. [42][43][44][45][46][47] However, a majority of this work was done to study the addition of a negative added bias on either the powered or grounded electrodes, rather than a positively added bias. While these works are useful for gaining a general understanding of the effects of dc bias on the discharge, the complete picture for the purposes of SRF cavity etching remains incomplete.…”