Chapter 19 - Data Analysis and Process Control

Abstract

Data analysis, model design, and control strategy are all important in utilizing the information from process sensors. Analysis of the raw data is relatively straightforward in fundamental and process-development studies, but can become overwhelming during manufacturing when continuous readings from the multiple sensors in each tool start streaming in for realtime analysis. Models are needed to transform these data into an understanding of the process (from a fundamental or process-development viewpoint) or at least a characterization that interrelates the state of the equipment, the process, and the wafer from a manufacturing perspective. These models are mathematical constructs that need not necessarily be based on the physical and chemical steps, and they can be important both in process development and control. Some models Utilize sensor readings of intermediate process conditions {in situ measurements), while others relate input process variables (equipment-state inputs) to process results (final wafer state).