Word Line Short Detection in 3D Memory via BVC ROI Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In 3D memory array manufacturing, detecting word line shorts in contact through-holes is challenging due to the rapid change in CT distribution, making it difficult to accurately identify defective positions using existing electronic beam inspection methods, which are time-consuming and lack precision.
Innovation Solution
A method and device that determine a region of interest in bright voltage contrast images, calculate coordinates, and review grayscale values to differentiate between charged and uncharged contact through-holes, automatically identifying defective contacts and displaying them for further processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If electronic beam inspection is used to detect leakage on CTs, then detection capability is provided, but inspection time becomes excessively long (half a day) and accuracy deteriorates due to difficulty in matching BVC positions with CT positions
Solution Approach 1:
The patent divides the inspection process into distinct segments: first determining the ROI region containing charged word lines, then identifying CT positions within this segmented region, and finally analyzing grayscale values at these coordinates. This segmentation reduces the complexity of matching hundreds of thousands of BVC positions with CT positions across the entire wafer, thereby reducing inspection time while maintaining accuracy.
Solution Approach 2:
The patent performs preliminary actions by first determining the ROI region and pre-identifying the coordinates of charged word lines before conducting the actual defect detection. This preliminary processing eliminates the need to scan the entire wafer surface, significantly reducing inspection time while ensuring accurate detection by focusing only on relevant regions.
2Reliability
If multiple CT positions are inspected to detect leakage, then comprehensive detection is achieved, but the rapid change in CT distribution makes it difficult to match BVC positions with corresponding CT positions
Solution Approach 1:
The patent applies local quality by focusing the analysis on the ROI region where charged word lines are located, rather than uniformly processing the entire wafer. Within this local region, the patent specifically analyzes the grayscale values at pre-determined CT coordinates, making the detection process tailored to the specific characteristics of charged word lines and their corresponding CT positions, thereby reducing position matching difficulty while maintaining comprehensive detection reliability.
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 reduces inspection time from half a day to under three minutes, enhances accuracy to over 98%, and provides tier analysis and wafer-level mapping insights, improving the detection of word line shorts in 3D memory arrays.
Implementation Method 1
multiple previously described CT positions are inspected by an electronic beam (E-Beam) inspection machine to detect leakage on the CTs
Implementation Method 2
bright voltage contrast (BVC) image for contact through-holes (CTs) of word lines in the memory device, the CTs comprising first CTs in at least one first word line that is electronically charged
Data Source
AI summary
A method for detecting word line shorts in a memory device includes: determining a region of interest (ROI) region in a bright voltage contrast (BVC) image for contact through-holes (CTs) of word lines in the memory device, the CTs comprising first CTs in at least one first word line that is electronically charged and second CTs in at least one second word line that is not electronically charged, and the ROI region comprising the at least one first word line; determining coordinates of each CT in the ROI region; reviewing a grayscale value at a position of the coordinates of each CT in the ROI region to determine whether the CT is a defective CT; and displaying the defective CT in response to the defective CT being determined.