Enhancing Process Engineering Decision-Making with Model Visualization Techniques

This paper examines the effectiveness of an interactive GUI for recipe adjustment in the PE-CVD process. By visualizing the relationship between film characteristics and parameters, we demonstrate that engineers can efficiently adjust recipes using their domain knowledge, even without expertise in data science. The experiment involved 41 batches of film deposition, and the GUI enabled the quantification of the impact of parameter variations on film properties. This approach is expected to enhance the efficiency and accuracy of process engineering.