System implementing machine learning in complex multivariate wafer processing equipment

Abstract

A system for controlling processing state of a plasma process is provided. One example system includes a plasma reactor having a plurality of tuning knobs for making settings to operational conditions of the plasma reactor. A plurality of sensors of the plasma reactor is included, where each of the plurality of sensors is configured to produce a data stream of information during operation of the plasma reactor for carrying out the plasma process. A controller of the plasma reactor is configured to execute a multivariate processing that is configured to use as input desired processing state values that define intended measurable conditions within a processing environment of the plasma reactor and identify current plasma processing values. The multivariate processing uses a machine learning engine that receives as inputs the desired processing state values and data streams from the plurality of sensors during processing of the plasma process. The machine learning engine is configured to identify current processing state values used to produce a compensation vector, such that the compensation vector defines differences between the desired process state values and the current processing state values. The controller is further configured to execute a compensation processing operation that transforms the compensation vector expressed in terms of measured conditions within the processing environment to changes of specific one or more of the tuning knobs of the plasma reactor.