SESHA 2013 Symposium Abstract

A Process-based and Simplified Carbon Footprint Model for Customized Semiconductor Products

Yin, Jessica; WANG, TING-HSIN
(UMC, Hsinchu, Taiwan)

Jessica Yin, Hsiao-Chun Wang, Tin Hsin Wang, Allen H. Hu, Ching-Yao Huang, Nowadays, “carbon footprint” is often used as shorthand for the amount of carbon emissions generated by an activity, organization or product. The carbon footprint label of a product has been the most important indication of how “green” a product is, especially as climate change effects increase in severity and consumer green awareness grows. The foundation to obtain a green label is to confirm data thoroughly during the inventory stage. For the semiconductor industry, the complicated manufacturing process of IC (integrated circuits) makes carbon-related basis data more difficult to inventory. The significant amount of manpower, time and budget involved also means that large enterprises are the few that have the means to conduct and market green measurement activities. In addition, the absence of a means to distinguish between different products is problematic. The current IC PCR announced in 2008 is a popular structure that has been used for calculating carbon footprint within the semiconductor industry. Following the IC PCR, we can get a rough result of the carbon footprint for each IC, only by averaging all of the products in the whole factory. The calculation is unable to show the difference between products with different process or raw material usage and cannot be applied to the design stage of a product. This study aims to develop a simplified model and a cost efficient, easy-to-use tool to meet semiconductor industry needs for carbon footprint calculation. We also developed a method which could differentiate the carbon footprint of various IC products by looking into the process design, tool and raw material usage. In total, we obtained more than 7,000 CFP samples from 8 semiconductor fabs from 2010 to 2011. Then we used regression analysis to find the relationship between key factors in the process that customers are concerned about and the carbon footprint of each product. Based on the results of the analysis, we identified 7 key parameters that predominantly affect on the carbon footprint of an IC. The key parameters include four for the front-end processes: function, technology node, mask layer and metal layer, and three for back-end process: packaging type, wire used and body size. The simplified carbon footprint model developed for those factors will help IC designers or even end product designers easily estimate carbon footprint in the design stage within a certain reliability, and help them evaluate if any changes to their design are necessary. Keywords: Carbon footprint of products (CFP), Green design, Semiconductor industry,