Computer-assisted measurement method for semiconductor manufacturing machine

摘要

The method involves determining various measurement data for a number of measurands. Based on the measurement data, the measurand is allocated a priority that indicates the significance of the measurand to the semiconductor manufacturing process. Measurands with the highest priority are identified and selected. To produce the priority, a neural network is used. Training data may be generated by classifying measurement data using a suitable semiconductor analysis method into perfect and imperfect processes. The neural network learns to distinguish between perfect and imperfect processes.