Diagnostic method and apparatus of the processing system using adaptive multivariate analysis

摘要

The present invention relates to a method and apparatus for monitoring a processing system for processing a substrate for semiconductor manufacturing processes. Only that, the data must be acquired from the processing system for a plurality of times of observation (step 510), the data including a plurality of data parameters. The main component analysis (PCA: Principal Components Analysis) model is constructed from the data, and a centering factor (step 520). The additional data is acquired from the processing system (step 550), the additional data comprises a plurality of further observations of data parameters. The centering factor is adjusted to yield an updated adaptive centering coefficient data for each parameter of the PCA model. The updated adaptive centering coefficients are applied to respective data on the PCA model parameters (step 560). At least one statistic is determined from the additional data using the PCA model (step 560). The control limits are set for statistic (step 570), it is compared with a statistic (step 580).