Development of a cleaning endpoint evaluation sensor to reliably monitor plasma chamber cleaning.

In this study, a cleaning endpoint evaluation sensor (CEES) was developed and applied to improve the monitoring accuracy and non-uniformity control in the cleaning processes for NF3-based remote plasma sources in semiconductor equipment. Although an increase in the chamber pressure increased the density of fluorine radicals and improved the overall etch reactivity, significant cleaning nonuniformity was observed in structurally limited regions, such as the substrate bottom, where the reactive species failed to reach sufficiently. Conventional diagnostic methods such as optical emission spectroscopy (OES) are limited in their ability to assess cleaning completion in such dead volume regions. In contrast, the proposed CEES directly measured the removal status of thin films mimicking actual process residues, providing real-time and spatially resolved endpoint evaluations. The CEES showed a strong correlation with the actual etch behavior and successfully identified regions where cleaning was incomplete, despite OES-based endpoint detection. The results revealed that the CEES can serve as an effective tool for monitoring cleaning non-uniformity and determining the true endpoint of plasma-cleaning processes. This approach has the potential for integration into advanced diagnostic platforms to enhance the reliability and efficiency of semiconductor manufacturing.