Weighted Pixel Defect Detection for Semiconductor Substrates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current defect detection methods for semiconductor substrates are inadequate in accurately identifying defects, particularly when comparing patterns with varying distances from lighting, leading to inefficiencies in the semiconductor fabrication process.
Innovation Solution
A device and method that utilize an image pickup unit with pixels to generate substrate images, a controller to detect defects by comparing weighted pixel values of pattern images, and a comparator to calculate difference values or normalized weighted values, allowing for precise defect detection based on threshold values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional defect detection methods are used to compare patterns, then the detection process is simple, but the detection accuracy deteriorates when patterns have different distances from lighting
Solution Approach 1:
The patent applies parameter changes by transforming pixel values into weighted values using importance weights. The controller calculates weighted values by multiplying pixel values by their corresponding importance weights, and then compares these weighted values instead of raw pixel values. This parameter transformation enables accurate defect detection even when patterns are at different distances from the light source, as the weighting compensates for illumination variations.
Solution Approach 2:
The patent introduces an intermediary element - the importance weight - that mediates between the pixel values and the comparison result. The weight acts as a mediator that adjusts the influence of each pixel value based on its reliability, which is determined by factors such as distance from the light source. This intermediary allows the system to handle varying lighting conditions without requiring complex hardware modifications.
2Reliability
If patterns at different distances from lighting are compared using conventional methods, then the process is fast, but the detection reliability deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing importance weights for each pixel position before the actual defect detection process. The controller determines the importance weight based on factors such as distance from the light source and stores these weights in advance. During defect detection, the system only needs to retrieve these pre-calculated weights and perform simple multiplication with pixel values, rather than performing complex calculations in real-time, thus maintaining fast processing speed while improving reliability.
3Ease of manufacture
If simple pixel value comparison is used, then the detection method is easy to implement, but the ability to detect defects in patterns with varying lighting conditions deteriorates
Solution Approach 1:
The patent transforms the simple pixel value comparison into a weighted value comparison by introducing importance weights as a new parameter. The controller multiplies each pixel value by its corresponding importance weight, which reflects the reliability of that pixel based on lighting conditions. This parameter change maintains the simplicity of the comparison operation while significantly improving adaptability to varying lighting conditions, as the weighting automatically compensates for illumination differences without requiring complex image processing algorithms.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables more accurate detection of defects on semiconductor substrates by comparing weighted or normalized pixel values, improving the reliability and efficiency of the defect detection process.
Implementation Method 1
an image pickup unit including pixels, the image pickup unit being configured to generate a substrate image by picking up an image of a substrate
Data Source
AI summary
In a device for detecting a defect, the device includes: an image pickup unit including pixels, the image pickup unit generating a substrate image by picking up an image of a substrate having patterns formed on a top surface thereof; and a controller for detecting a defect located on the substrate, based on the substrate image, wherein the substrate image includes pattern images corresponding to the patterns, wherein each of the pattern images includes pixel values, wherein the controller detects the defect by comparing weights of pixel values for each of the pattern images.