Accelerated defectivity monitoring in high-volume manufacturing via unsupervised AI techniques
We present an innovative framework that leverages novel unsupervised AI techniques to automate defectivity setup in IC chip manufacturing. This new approach minimizes the need for expert intervention and eliminates redundant inspector-SEM review loops. With shorter time-to-results, the Fab productivity during the ramp-up phase can be greatly improved. This modern framework ultimately helps in bringing the industry closer to the vision of fully automated factory operation, allowing end users to focus more effectively on yield analysis.
Authors
- Paolo Parisi
- C. Régnier
- A. S. Menon (ORCID: https://orcid.org/0009-0003-3911-1744)
- Daniel Arias
- A. Grimaud
- Brian Ries
- Aditya Tiwari
- V. Ramakrishna (ORCID: https://orcid.org/0009-0003-0228-7674)
- Parikshit Pratim Sharma (ORCID: https://orcid.org/0000-0001-7353-3690)
- Guillaume Celi
- Hugo Ouvrard
- Pranay Deevanapelli
- Hanoch Kopel
- Marie-Laure Pasqualini
- Sumant Anurag
- Sayi Tummalapalli
Institutions
- Centre Lillois d'Etudes et de Recherches Sociologiques et Economiques (FR)
- STMicroelectronics (France) (FR)
- KLA (United States) (US)
Publication Details
- Published
- 2025-05-05
- DOI
- https://doi.org/10.1109/asmc64512.2025.11010333
- Primary Topic
- Industrial Vision Systems and Defect Detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00