Advanced CMP Process Control by Using Machine Learning Image Analysis

Recent Progress on Optical Tomographic Technology for Measurements and Inspections of Film Structures

“…Unlike the traditional measurement strategy to analyze the measurement results based on the theory, recently, machine learning (ML) or deep learning algorithms have been applied to film metrology with the huge computing power [ 107 , 108 , 109 , 110 , 111 , 112 ]. In various science and engineering fields, ML algorithms are being increasingly used to analyze characterization data, although the physical or analytical model is not completely found out.…”

“…Instead, lots of experimental data sets to train the algorithm should be prepared. Several research works based on ML algorithms have been recently reported, and they measured and monitored the growth of the film layers during the film deposition process [ 108 , 109 ], film thickness variations during the CMP process [ 110 ] and the optical property of materials [ 111 ]. The ML-based film metrology was recently extended to the determination of a multi-layered wafer stack structure in semiconductor applications [ 112 ], as shown in Figure 14 , and each layer thickness was able to be non-destructively determined with an average of approximately 1.6 Å as RMS error.…”

“…Unlike the traditional measurement strategy to analyze the measurement results based on the theory, recently, machine learning (ML) or deep learning algorithms have been applied to film metrology with the huge computing power [107][108][109][110][111][112]. In various sci-) indicates the relative performance value.…”