Run-to-Run Control Automation in CMP Enabled by a Production-Oriented Digital Twin
Run-to-run (R2R) process control is a core operational practice in semiconductor manufacturing, relying heavily on continuous monitoring, model tuning, and engineer intervention to maintain stable process performance. In CMP processes, these challenges are particularly pronounced due to strong interactions among consumables, equipment conditions, and inter-run variability, making consistent R2R operation difficult to sustain in practice.
This presentation introduces a practical framework for run-to-run control automation in CMP, leveraging digital twin modeling to support adaptive R2R operations. The approach brings together modeling, virtual metrology, and control-related capabilities within a unified operational context that reflects how R2R control is managed on the fab floor. The framework emphasizes operational alignment and stability across runs, helping reduce variability and improve control consistency in complex CMP environments.
The proposed framework is designed to coexist with existing APC approaches, operating as a complementary layer that supports R2R control activities while preserving established fab workflows and operational ownership. By focusing on CMP as a representative and high-impact starting point, this work outlines a realistic direction for enhancing run-to-run control robustness without expanding scope beyond what can be practically adopted and maintained in semiconductor fabs.