A Bayesian indicator for Run-to-Run performance assessment in semiconductor manufacturing

In this paper, Bayesian theory is used to build an indicator for Run-to-Run control. This indicator is used for assessing the performances of the regulation loops in a batch industry. The indicator is using four main inputs which are the output/target error, the dispersion of the output, the out of tolerance rate (oot) and the value of the industrial risk. The efficiency of the proposed Bayesian method has been tested on the deposition area of a semiconductor foundry.