Optical Emission Spectroscopy Analysis of the BOSCH Process Utilizing k-Means Clustering and Dynamic Mode Decomposition

Plasma etching processes are an integral element in semiconductor manufacturing. This study focuses on the BOSCH process, which is distinguished by multiple and quick process shifts. Optical emission spectroscopy (OES) is used to monitor information about the plasma without interfering with it. By monitoring the chemical changes across different phases of the process, it is possible to evaluate process and chamber information. This investigation utilizes k-means and dynamic mode decomposition (DMD) to analyze the OES data. K-means is a clustering algorithm that separates signals into a predetermined number of clusters, while DMD is a dimensionality reduction technique.