Authors: D. Ortloff, J. Popp
Affilation: Process Relations GmbH, Germany
Pages: 741 - 744
Keywords: statistical experiment design, process development, fractional factorial design
Process development, especially in the high tech sectors, is an extremely complex task. The key challenge is developing a range of process recipes and testing these recipes via experimentation. These experiments can take a lot of time and resources and the retrieved results are often not conclusive so that further experiments need to be defined. Over the past twenty years, there has been a focus on shortening development cycle times. Consequently there has been increasing use of various statistical tools such as ‘Statistical Experiment Design’ (SED) designed to help automate and accelerate the design process. Statistical experimental design, especially Fractional Factorial Design (FFD), is a method to significantly enhance experimentation setup and execution. FFD acts as a filter for experiment designs and recipes and uses statistics to expose information about the most important features of the problem studied, while using a fraction of the effort of a full factorial design in terms of experimental runs and resources. However, there is a major problem with FFD - it is more vulnerable and sensitive to errors. This paper will introduce a software driven development approach extending and maturing the current practices of SED.