Connecting process control systems and data to advance AI technology

As semiconductor technology scaling moves to the atomic era, device complexity continues to increase. New process technologies (high NA EUV lithography), new integration schemes (advanced packaging, multi-die heterogenous integration), novel device architectures (CFET, backside power delivery) and shrinking pitches cause higher variability and tighter process margins. To address these production challenges first requires a comprehensive inspection, metrology and data analytics portfolio that supports modeling, characterization and monitoring from design to reticle to process modules to packaging. Multi-data streams then feed ML-based engines that take the performance of these systems to the next level by distilling multi-variate factors into meaningful response prediction, dynamic sampling and optimized process control. These process control innovations drive production quality, reliability and yield and support the industry’s advancements in AI chip technologies.