Semiconductor Pattern Inspection Using Multi-Directional Width Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for inspecting semiconductor chip patterns are inefficient due to the need for numerous threshold values to account for varying pattern density across the chip, leading to prolonged setup times and reduced working efficiency.
Innovation Solution
A method that measures pattern widths in multiple directions, compares these measurements to determine the shortest width, and evaluates defects based on this information, allowing for automatic detection and evaluation of defects in semiconductor chip patterns, regardless of their complexity or orientation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple threshold values are set according to positions in the pattern to account for varying pattern density, then measurement precision is improved, but device complexity increases and working efficiency decreases
Solution Approach 1:
The patent changes the parameter from multiple position-specific threshold values to a single threshold value applied uniformly across the entire pattern. This is achieved by transforming the original image data through filtering operations that normalize the varying pattern density, allowing one threshold to effectively detect defects throughout the whole pattern regardless of local density variations.
Solution Approach 2:
The patent creates a universal threshold value that serves all positions in the pattern simultaneously, rather than having position-specific thresholds. The filtering process generates transformed image data where a single threshold can universally detect defects across regions with different pattern densities, making the inspection system more versatile and easier to operate.
2Measurement precision
If multiple threshold values are set according to positions in the pattern to account for varying pattern density, then measurement precision is improved, but productivity decreases
Solution Approach 1:
The patent transforms the inspection parameter from multiple position-dependent thresholds to a single universal threshold by applying filtering operations to the image data. This parameter transformation maintains defect detection accuracy across varying pattern densities while eliminating the time-consuming process of setting multiple thresholds, thereby improving productivity.
Solution Approach 2:
The patent performs preliminary filtering of the image data before defect detection, transforming the original image into filtered image data that normalizes pattern density variations. This preliminary action prepares the data in advance, allowing subsequent defect detection to use a single threshold value without requiring time-consuming manual threshold setting for each position, thus improving working efficiency.
3Measurement precision
If the pattern is finely segmented to account for density variations, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The patent replaces the manual mechanical process of setting multiple thresholds for different positions with an automated image filtering process. The filtering operation automatically transforms the image data to account for pattern density variations, eliminating the need for manual threshold setting and segmenting, thereby reducing time loss while maintaining measurement precision.
Solution Approach 2:
The patent changes the approach from manual parameter setting (multiple thresholds) to automated parameter transformation (image filtering). The filtering process automatically adjusts the image data to compensate for density variations, replacing the time-consuming manual threshold setting process with an automated computational approach that maintains accuracy without the time penalty.
Data Source
AI summary
A method for inspecting a pattern includes measuring, in a first direction, a width of a reference pattern at plural positions in the reference patter; measuring, in a second direction, a width of the reference pattern at the plural positions. Comparing the first and second width and determining which of the first and second widths is shortest; extracting a defect in a pattern to be inspected; and evaluating the extracted defect depending on the determined direction.