Comparison of CHARM-2 and surface potential measurement to monitor plasma induced gate oxide damage

Plasma process induced gate oxide damage was found in early process development stages. Device data showed unacceptable burn-in failure. By utilizing multiple test vehicles, the underlying cause of the oxide damage was identified. This study showed that no single methodology is adequate for damage monitoring. A combination of monitoring techniques is required to understand the root cause of the damage and how to optimize the process or equipment. The plasma process was optimized and verified with CHARM-2 monitor wafer response. Further device data verification indicated that no gate oxide damage was found with the improved process. A fast turnaround time for the plasma monitors was essential to understand and determine the plasma damage source. Understanding the relationship between plasma monitor response and plasma process is a key point to identify the damage source. A fingerprint of the plasma process is very useful for process control and defect reduction.