Data-driven plasma modelling: Surrogate collisional radiative models of fluorocarbon plasmas from deep generative autoencoders

TL;DR

This paper presents a deep generative autoencoder model that accurately predicts optical spectra and images of plasma based on input parameters, enabling fast simulations and aiding plasma research.

Contribution

The authors developed and released a deep autoencoder-based surrogate model trained on a large dataset to generate plasma spectra and images from input parameters, facilitating rapid predictions.

Findings

Model generates spectra and images in just over 10 seconds on a GPU.

Autoencoder trained on 812,500 data pairs across various plasma gases.

The approach enables fast, data-driven plasma modeling and is accessible via Google Colab.

Abstract

We have developed a deep generative model that can produce accurate optical emission spectra and colour images of an ICP plasma using only the applied coil power, electrode power, pressure and gas flows as inputs -- essentially an empirical surrogate collisional radiative model. An autoencoder was trained on a dataset of 812,500 image/spectra pairs in argon, oxygen, Ar/O\textsubscript{2}, CF\textsubscript{4}/O\textsubscript{2} and SF\textsubscript{6}/O\textsubscript{2} plasmas in an industrial plasma etch tool, taken across the entire operating space of the tool. The autoencoder learns to encode the input data into a compressed latent representation and then decode it back to a reconstruction of the data. We learn to map the plasma tool's inputs to the latent space and use the decoder to create a generative model. The model is very fast, taking just over 10 s to generate 10,000 measurements on a single GPU. This type of model can become a building block for a wide range of experiments and simulations. To aid this, we have released the underlying dataset of 812,500 image/spectra pairs used to train the model, the trained models and the model code for the community to accelerate the development and use of this exciting area of deep learning. Anyone can try the model, for free, on Google Colab.