Dense Pattern Defect Detection for Rare EUV Stochastic Defects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current EUV Lithography techniques are inefficient in detecting rare stochastic defects of nanometric scale, which are too small to be detected by optical tools and require time-consuming scanning of the entire substrate using charged particle beams.
Innovation Solution
The implementation of a system and method that uses dense patterns on a substrate, which increases the chances of detecting rare stochastic defects by scanning a small fraction of the substrate with charged particle tools, allowing for a reasonable detection time, involving steps to search, estimate, and respond to defect occurrences based on defect density thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If charged particle beam tools scan the entire substrate to detect rare stochastic defects, then detection reliability is improved, but detection time increases significantly
Solution Approach 1:
The substrate is divided into multiple regions of interest (ROIs) based on process parameters and defect probability maps. Instead of scanning the entire substrate, the charged particle beam tool only scans these segmented high-risk regions, reducing detection time while maintaining reliability for rare stochastic defects.
Solution Approach 2:
The patent applies partial action by scanning only a fraction (e.g., 1-10%) of the total substrate area that corresponds to regions with higher defect probability. This partial scanning approach is sufficient to detect rare stochastic defects while avoiding the time cost of full-substrate scanning.
2Productivity
If optical tools are used to detect defects, then detection speed is improved, but detection precision for nanometric defects deteriorates
Solution Approach 1:
The patent applies local quality by using optical tools for initial fast screening of large substrate areas, then switching to charged particle beam tools for precise inspection of specific regions of interest where rare stochastic defects are most likely to occur. This combines the speed of optical tools with the precision of charged particle tools at critical locations.
Solution Approach 2:
The patent replaces optical detection (electromagnetic field-based) with charged particle beam detection (particle-based) for the specific task of detecting rare stochastic defects in regions of interest. This substitution enables nanometric-scale precision while maintaining practical detection speeds through targeted scanning.
3Manufacturing precision
If the defect density threshold is set low to ensure quality, then product quality is improved, but the number of rejected substrates increases
Solution Approach 1:
The patent uses feedback from process parameters (exposure dose, focus, resist coating uniformity) and preliminary optical inspection results to dynamically adjust defect density thresholds and identify regions of interest. This feedback mechanism allows for optimized quality control that reduces false rejections while maintaining high manufacturing precision for rare stochastic defects.
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 significantly increases the probability of detecting rare stochastic defects by a factor of 10^8, enabling efficient and reliable detection within a practical time frame, thereby improving the quality control of semiconductor wafers.
Implementation Method 1
searching for a rare stochastic defect in a dense pattern of a substrate by illuminating the dense pattern with a charged particle beam and generating images of the dense pattern
Data Source
AI summary
A method for detecting a rare stochastic defect, the method may include searching for a rare stochastic defect in a dense pattern of a substrate, wherein the rare stochastic defect is (a) of nanometric scale, (b) appears in a functional pattern of the substrate with a defect density that is below 10−9, and (c) appears in the dense pattern with a defect density that is above 10−7; wherein the dense pattern is a dense representation of the functional pattern that differs from the functional pattern by at least one out of (a) a distance between features of the dense pattern, and (b) a width of the features of the dense pattern; and estimating the occurrence of the rare stochastic defect within the functional pattern based on an outcome of the searching.