Runtime PRNU Correction for Tilted Wafer Inspection Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Semiconductor inspection systems face luminance inaccuracies due to photo response non-uniformity (PRNU) caused by substrate contours such as tilt or warp, which affect image sensors' uniformity and lead to pixel-level luminance inconsistencies.
Innovation Solution
A method to dynamically correct PRNU at runtime by determining the substrate's contour using a semiconductor substrate inspection system, generating a correction model based on the relationship between grayscale values and angles of incidence, and applying this model to images to adjust for substrate tilt and warp.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If PRNU correction is applied using standard calibration, then image uniformity is improved, but the correction becomes inaccurate when substrate contour (tilt or warp) is present
Solution Approach 1:
The system performs preliminary actions by determining the substrate contour and calculating the angle of incidence before applying PRNU correction. This allows the correction to be adapted to the actual substrate orientation, ensuring both image uniformity and correction accuracy even when the substrate is tilted or warped.
Solution Approach 2:
The system changes the correction parameters based on the determined angle of incidence. By adjusting the PRNU correction according to the substrate's angular orientation, the system maintains measurement precision while accommodating variations in substrate contour that would otherwise compromise reliability.
2Productivity
If PRNU correction is applied without considering substrate contour, then processing speed is maintained, but luminance inaccuracy increases
Solution Approach 1:
The system performs a preliminary contour determination and angle calculation that can be executed efficiently before the main inspection process. This preliminary action enables subsequent PRNU correction to be applied with high luminance accuracy without significantly impacting overall processing speed.
Solution Approach 2:
The system replaces complex mechanical substrate positioning with computational methods. By using image analysis and mathematical calculations to determine contour and angle of incidence, the system achieves high luminance accuracy through software-based correction rather than mechanical adjustment, maintaining processing speed.
3Measurement precision
If dynamic PRNU correction at runtime is implemented, then luminance accuracy is improved, but system complexity increases
Solution Approach 1:
The system performs self-service by automatically determining its own operational parameters (angle of incidence) from the captured images and autonomously adjusting the PRNU correction accordingly. This eliminates the need for external manual calibration or complex additional hardware, achieving high luminance accuracy while minimizing system complexity.
Solution Approach 2:
The system implements feedback by using the captured images to determine substrate contour and angle of incidence, then applying this information to adjust the PRNU correction in real-time. This closed-loop approach improves luminance accuracy through dynamic adaptation without requiring complex external control systems.
Data Source
AI summary
Image sensors assist in inspecting semiconductor substrates, including semiconductor wafers, during the semiconductor manufacturing process. However, due to the non-uniformity of cells between image sensors, which is referred to as photo response non-uniformity (PRNU), each image sensor can generate a different voltage level when illuminated with a uniform light source, which leads to a luminance inaccuracy at the pixel level that is referred to as photo response non-uniformity (or illumination flat fielding). Photo response non-uniformity (PRNU) can be corrected using a system calibration that calibrates to a substrate, including wafers, for a given illumination condition, objective, and camera. The PRNU correction is then applied to all images acquired with the given setup to produce a uniform image. Techniques are described to correct the PRNU at runtime even when there is a contour, e.g., tilt or warp, in the substrate as it relates to the normal axis of the optics.