Virtual IED sensor for df rf CCP discharges

TL;DR

This paper presents a virtual IED sensor for real-time plasma process control in rf CCP discharges, using a fast calculation method validated across different gases, enabling accurate IED estimation without direct measurement.

Contribution

It introduces a novel virtual IED sensor based on a fast calculation method incorporating sheath waveform data and Monte Carlo simulations, suitable for real-time plasma processing.

Findings

Adequate IED estimation achieved across various gases.

Estimation time is approximately tens of seconds on modern PCs.

Validation confirms the sensor's effectiveness in real plasma conditions.

Abstract

Ion-assisted surface processes are the basis of modern plasma processing. Ion energy distribution (IED) control is critical for precise material modification, especially in atomic-level technologies such as atomic layer etching. Since this control should be done in real time, it requires real-time feedback using fast process sensors. In the general case of an industrial plasma reactor, when direct IED measurement is not possible, the IED can be estimated using the concept of a virtual IED sensor. In this paper, a similar virtual IED sensor is considered using an asymmetric dual-frequency (df) rf CCP discharge as an example. It is based on a fast calculation method of the IED at an rf-biased electrode. This approach uses the experimentally measured sheath voltage waveform and plasma density (or ion flux) as input data, and also includes Monte-Carlo simulation of ion motion in the sheath to take into account the effect of ion-neutral collisions. To validate this approach, experiments were carried out using various plasma diagnostics in several gases: argon and xenon as examples of plasma with atomic ions and nitrogen as an example of plasma with molecular ions. It is shown that in all cases it is possible to obtain an adequate IED estimation, close to the experimental one, in a reasonably short time (~ tens of seconds when using a modern PC). The results obtained demonstrate the possibility of using a virtual IED sensor in real plasma processing.