Directly addressing uncertainty in ESH evaluation

Two of the major barriers for integrating comprehensive environmental evaluation methods such as the life cycle analysis (LCA) into technology selection in the semiconductor industry are: (1) due to the large amount of data requirement, the time needed for a in-depth LCA does not match with the fast pace of innovation cycle of the industry; (2) with new materials constantly being introduced into testing and production, environmental, safety, and health data of these new material are often lacking. To address these two barriers, it is proposed that analyses should start with simplified models and low-quality data, and then based on the uncertainty range of the analysis results, the decision maker decide on whether to refine the models and collect more data, and which input factors to refine. To enable this, the uncertainty in the results and the contributions of the inputs to this uncertainty need to be quantified. This paper describes the uncertainty analysis and sensitivity methods that can be used for the above purpose. A case study of comparing two chamber cleaning gases is used to illustrate the methods.