Development of Advanced Data Analytics Software to Optimise the Yield and Reduce Yield Variations in Advanced 300 mm Wafer Foundry
GlobalFoundries Singapore Pte Ltd (“GF”) and A*STAR I²R entered into a Joint Collaboration Agreement to develop a customized Yield Analysis Software to optimize Yield and reduce Yield Variability for its advanced 300mm Operations. This will be achieved using A*STAR I²R’s state-of-the-art Data Analytics techniques/tools, jointly developed with GlobalFoundries Singapore Pte Ltd.
Across GF’s extensive range of products, the operational focus of maintaining sustainable yield is paramount. Reducing yield variations across and within lots and wafers are equally important as achieving high and predictable yields.
In a highly complex wafer manufacturing facility, the manufacturing process can involve a large number of steps with extensive equipment, process, metrology, consumables, electrical test, operational and environmental measurements.
Such multidimensional, though structured data posed a challenge in determining the factors affecting the wafer yield and the variations in yield. Coupled with the low amount of data of interest of both low yield and high yield variability outlier wafers, the development of accurate, sustainable and scalable yield analysis using the correlation in data and measurements will require both deep data analytics expertise and savvy domain knowledge.
Utilising A*STAR I²R’s data analytics techniques/tools such as feature engineering, data re-balancing, correlation analysis, significance testing, sensitivity analysis, and modelling, a customised yield analysis workflow is developed, to determine the factors causing the variations of yield in GF’s range of products.
This yield analysis workflow will also investigate and identify the optimal setting range for key factors leading to higher yield and lower yield variation of products at the various stages of the product development cycle.