Defect Detection on Semiconductor Wafers by Distribution Analysis

A method for object classification that is based on distribution analysis is proposed. In addition, a method for finding relevant features and the unification of this algorithm with another classification algorithm is proposed. The presented classification algorithm has been applied successfully to real-world measurement data from wafer fabrication of close to hundred thousand chips of several product types. The presented algorithm prefers finding the best rater in a low-dimensional search space over finding a good rater in a high-dimensional search space. Our approach is interesting in that it is fast (quasi-linear) and reached good to excellent prediction or detection quality for real-world wafer data.

Fast Accurate Defect Detection in Wafer Fabrication

A generic fast method for object classification is proposed. In addition, a method for dimensional reduction is presented. The presented algorithms have been applied to real-world data from chip fabrication successfully to the task of predicting defect states of tens of thousands of chips of several products based on measurements or even just part of measurements. Unlike typical neural networks with a large number of weights to optimize over, the presented algorithm tries optimizing only over a very small number of variables in order to increase chances to find a global optimum. Our approach is interesting in that it is fast, led to good to very good performance with real-world wafer data, allows for short implementations and computes values which have a clear meaning easy to explain.