Abstract
C2C (chamber-to-chamber) matching of the plasma processing reactors, called tool-to-tool matching, is one of the most important procedures to perform mass production in the organic light emitting diode display or semiconductor manufacturing facilities. Plasma processes, which adopt the RF power, are governed by the generation rate of reactive radicals and ions and their transportation to the reacting surfaces. Therefore, the characteristics of plasma heating, which determine the energy gain and relaxation of the electrons, are important information for performing C2C matching in the electron impact collisional inelastic processes dominated plasmas. In this study, we introduced the plasma heating characterizing PI (plasma information) index by using the limited sensor data accessible in the mass-producing fab with the help of a simple machine learning methodology to extract the information about the plasma heating properties. This index is applied to the C2C matching of the large-area capacitively coupled RF discharge-based PECVD (plasma-enhanced chemical vapor deposition) reactors successfully. In addition to the PI parameterization of the plasma heating characteristics in the large-area inductively coupled discharge applied plasma etcher introduced in the previous study, we could introduce the plasma heating characterized PI index for the PECVD reactors. From this, we could validate its utility to the artificial intelligence modeling, such as the plasma information-based virtual metrology and plasma information-based advanced process control, to optimize the productivity of the mass-producing factory.