US20250308960
2025-10-02
Electricity
H01L21/67253
The application discusses a system designed to detect and monitor objects, such as bubbles, within a fluid bath used in semiconductor processing. By analyzing video data captured over time, the system tracks these objects to gather metrics like their position and size. This data is crucial for correlating the parameters of processing recipes with the quality of the resulting wafers, allowing for adjustments to improve manufacturing outcomes.
The system employs processing circuitry that analyzes video frames to detect objects in a fluid. It calculates metrics for these objects across multiple frames, determining changes in their properties such as position and size. This analysis helps identify whether objects in different frames are the same, facilitating accurate tracking over time. The method also involves assigning probability values to assess object continuity between frames.
Key components include a light source and an imaging device. The light source emits light at specific wavelengths through the fluid to enhance object visibility, while the imaging device captures this data. The processing circuitry then applies filters to video frames, transforming them for better object detection and metric determination.
The processing system uses the object metrics and wafer quality data to adjust processing recipes. By understanding how different recipe parameters affect wafer quality, the system can modify these parameters to optimize outcomes. This feedback loop ensures that the semiconductor manufacturing process continually improves based on real-time data.
This technology is particularly useful in semiconductor manufacturing where precise control over processing conditions is vital. By monitoring objects in fluid baths and adjusting recipes accordingly, manufacturers can achieve higher quality wafers with fewer defects. The approach provides a robust method for enhancing process reliability and efficiency.