The rapid innovation of new process technologies in the semiconductor industry along with continuously growing amounts of data result in increased challenges in the area of data management and analysis. Although computer analysis systems and statistical analysis methods provide invaluable assistance, intervention from experienced engineers is still required to interpret the results and proceed to other phases of analysis. This is due to the fact that complex relationships between large amounts of disparate data types often exist, and traditional computational methods do not automatically summarize these relationships adequately. Current data analysis methods consume large amounts of human resources in order to determine the root cause of process and yield excursions. High-quality analysis requires experienced engineers to evaluate the results, and the efficiency of an organization's ability to solve problems depends on retaining and communicating this expertise. Hence, it is important that a knowledge retention system be incorporated to improve the efficiency of root cause analysis. This article discusses the design of an automated multi-domain system for analyzing data, which develops and retains expert knowledge pertaining to semiconductor processing data. This analysis architecture determines the relationship among different data types and enables automated classification of the various correlations. In addition, this system provides a variety of different analysis modules, which can be chosen based on the particular problem to be solved. When coupled with the high-speed and high-volume processing capabilities of computers, this architecture captures expert knowledge, and improves analysis efficiency and accuracy.
展开▼