[16p-W9_222-1~11] Lab to Fab 2: Bridging Semiconductor Research with Scalable and Sustainable Manufacturing

3:50 PM - 4:25 PM JST(6:50 AM - 7:25 AM UTC)

[16p-W9_222-7] Enhancing semiconductor manufacturing technology through Process Informatics

〇Masaki Takaishi¹ (1.Aixtal)

Keywords:

Process Informatics,Semiconductor,Machine Learning

Semiconductor manufacturing must simultaneously achieve shorter development cycles and reduced environmental impact while improving quality and productivity. This talk introduces Process Informatics, a framework that integrates physics-based modeling with data-driven surrogate models to run virtual experiments on a digital twin and efficiently search for optimal process conditions via multi-objective and robust optimization. We present Aixtal case studies on (1) deposition tool design optimization, (2) robust optimization of wafer processing conditions, and (3) cross-process condition optimization, demonstrating improved exploration efficiency and reproducibility in decision-making. We also overview representative public examples and trends in (4) cross-process predictive anomaly detection and root-cause analysis, and (5) fab operations optimization including material-handling simulation and dispatching. Practical implementation considerations are discussed.