Advanced quality control for probe precision forming to empower virtual vertical integration for semiconductor manufacturing

• Published in: Computers & industrial engineering 2023-09, Vol.183, p.109461, Article 109461
• Main Authors: Fu, Wenhan, Chien, Chen-Fu, Chen, Chi-Hang
• Format: Article
• Language: English
• Subjects: Advanced quality control; Circuit probe; Key quality characteristics identification; Parameter optimization; Virtual vertical integration
• ISSN: 0360-8352; 1879-0550
• DOI: 10.1016/j.cie.2023.109461

Advanced quality control framework for precision forming based on experimental design, machine learning, and genetic algorithm is proposed to meet the specifications of multiple quality characteristics that are interrelated in light of changing production conditions for yield enhancement and production effectiveness to enhance the yield via virtual vertical integration among different segments of semiconductor industry. [Display omitted] •Novel approach is proposed for real-time parameter optimization to enhance data integrity of circuit probe test.•A decision support system is constructed for advanced quality control in precision forming.•Proposed approach effectively identifies key quality characteristics.•Parameter optimization solution is developed with the integration of DOE, PLS and GA.•An empirical study is conducted in semiconductor for validation. Circuit probe is crucial for electrically testing for functional defects to determine the known good dies before integrated circuit (IC) packaging. Semiconductor probe forming showing quality characteristics with highly correlated responses, yet little research has been done to develop effective solution for optimizing the parameters for probe precision forming. In practice, process parameters for probe forming need to be adjusted when some response variables become out of the specifications owing to degrading of production conditions, most companies rely on domain knowledge and trial-and-error approach. As IC critical dimensions are shrinking, the present problem for probe precision forming is increasingly challenging and critical for data integrity of circuit probe test for yield enhancement of IC products. Focusing on realistic needs, this study aims to develop an effective solution for real-time parameter optimization in light of changing production conditions based on design of experiment, partial least square, and genetic algorithm for advanced quality control to enhance data integrity of circuit probe test. The results of an empirical study have shown that the proposed approach can significantly enhance the effectiveness and efficiency of parameter optimization in real setting. Indeed, the developed solution has been implemented with significant advantages.