Abstract
Tools or chambers at a single step are designed to perform the same processing in semiconductor industry. In practice, tools or chambers differ and do not process lots identically. The ability to achieve consistent performance of wafer across the entire toolset is critical to developing and maintaining a high yield. In this paper, a statistical method is proposed to diagnose any reasonable difference between golden and inferior chambers that are classified in terms of end-of-line quality of wafer. Key features of the sensor-variable profiles are mined out to determine the causes of chamber mismatching in the manufacturing process. The method not only employs well-known statistical analysis techniques of discrimination and regression, but also presents a synopsis of analysis results in the chart of R 2 statistics versus p-value. This framework provides a systematic method of drawing inference from the available evidence without interrupting the normal process operation. The proposed concept is illustrated by an electroplating process in a local fabrication unit. © 2012 IEEE.