Virtual metrology-based engineering chain management by multi-classification of quality using support vector machine for semiconductor manufacturing

Abstract

This paper suggested that engineering chain management (ECM) that manages technical information and the process is necessary for the semiconductor industry. A virtual metrology-based (VM) manufacturing management system was proposed. This system realised the ECM concept, succeeded in real-time quality prediction and served a consistent driver to actions of all divisions of work. Support vector machine (SVM) was applied to construct an accurate VM model that provided multi-class quality prediction of the product. Although only the equipment variable data of a real mass production factory was used as an input, the VM model predicted with 100% accuracy the quality of the product.