Developing a parametric carbon footprinting tool for the semiconductor industry

Abstract

The advent of global awareness of sustainable development resulted in consumer demand for low-carbon electronic products. Thus, semiconductor industry has to develop additional core competencies in that direction to remain competitive. The present study aims to establish a parametric-based tool that can identify key parameters of the complicated manufacturing processes in the semiconductor industry to simplify the calculation of the carbon footprint of products (CFP). Six regression models for CFP were developed by applying process and statistical analyses on 7114 wafer products. Results indicate that these regression models with the three key parameters (mask layer, metal layer, and technology node) can effectively predict the CFP of wafer with six different function types because the adjusted R 2 of all regression models was >0.5. Moreover, the mask layer could be the most important parameter for predicting CFP of wafer because of its higher standardized coefficients (β) in each regression model. This methodology reduces the time, cost, and information requirements of the product in traditional life cycle assessment. The proposed method introduces criteria for low-carbon decision making in the semiconductor sector.