The application and validation of a new robust windowing method for the Poisson yield model

In this paper, a simple robust method of extracting the systematic and random components of yield is presented. The method has proven to be robust in its implementation in systems to automate the analysis of wafer probe bin map data spanning several years for multiple cleanrooms and technologies. The method's ability to detect systematic yield loss is evaluated and its dependence on die size is discussed. The application of data sub-setting to the results, which allows focused analysis of yield problems, is shown to be effective.