Lithographic Mask Design Verification via CD-SEM Contour Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lithographic mask design and inspection processes face challenges in accurately predicting and verifying the printed pattern, especially for complex geometries, due to limitations in measurement techniques like CD-SEM and PWQ, which struggle with sub-32nm node dimensions and complex 2D logic structures, leading to inaccuracies in Optical Proximity Correction (OPC) modeling.
Innovation Solution
A method for designing lithographic masks that involves simulating patterns using OPC-based methodologies, producing a real mask, and verifying the printed pattern by extracting experimental contours from CD-SEM measurements for direct comparison with intended and simulated contours, utilizing automated image processing to derive and compare key parameters such as line-end positions and widths, thereby improving accuracy and automation in the verification process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If CD-SEM measurement technique is used to measure critical dimension of pattern features, then measurement can be performed on printed dies, but only specific feature types can be measured such as regular line/space widths and distances between opposite line ends, excluding complex geometries
Solution Approach 1:
The patent creates a digital contour representation (copy) of the printed pattern from CD-SEM images. This contour model captures the complete geometry including complex features, while the physical CD-SEM measurement remains unchanged. The contour serves as an intermediate representation that preserves all geometric information without requiring manual measurement box placement.
Solution Approach 2:
The patent introduces contour extraction as an intermediary step between CD-SEM measurement and pattern verification. The contour acts as a mediator that translates complex 2D geometries into measurable parameters, enabling verification of features that were previously unmeasurable by standard CD-SEM techniques.
2Measurement precision
If manual measurement box placement is used in CD-SEM, then specific positions can be measured, but only a handful of positions are measured and the process is time-consuming
Solution Approach 1:
The patent implements automated contour extraction and parameter determination that performs measurements without human intervention. The system automatically identifies features, extracts contours, and calculates parameters such as line-end positions and widths, eliminating the need for manual measurement box placement and significantly increasing measurement throughput.
Solution Approach 2:
The patent performs preliminary contour extraction from the entire pattern area before parameter measurement. By pre-processing the image to create a complete contour map, all subsequent measurements can be performed automatically on extracted parameters rather than requiring manual intervention for each measurement point.
3Area of stationary object
If optical microscope technique is used for PWQ to scan all dies, then all dies can be examined, but the resolution is limited and does not allow to retrieve meaningful quantitative data for nm-level OPC modeling
Solution Approach 1:
The patent segments the measurement task into two stages: first, low-resolution optical scanning to identify die locations and overall pattern presence; second, high-resolution CD-SEM imaging of specific regions of interest for quantitative measurement. This segmentation allows both wide coverage and high precision to be achieved by using each technique for its optimal function.
4Manufacturing precision
If OPC modeling is performed at nm-level precision, then accurate prediction of printed pattern is achieved, but verification requires measurement techniques that can resolve such fine details, which current techniques cannot provide
Solution Approach 1:
The patent creates a digital contour copy from CD-SEM images that preserves nm-level geometric details. By working with the contour representation rather than direct physical measurements, the system can achieve verification precision that matches the nm-level OPC modeling accuracy, as the contour extraction process maintains sub-10nm resolution.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
The invention is related to a method for designing a lithographic mask to print a pattern of structural features, wherein an OPC-based methodology may be used for producing one or more simulated patterns as they would be printed through the optimized mask. A real mask is then produced according to the optimized design, and an actual print is made through the mask. To evaluate the printed pattern and the PW on wafer more accurately, experimental contours are extracted from the CD-SEM measurements of the printed pattern. The verification of the mask is based on a comparison between on the one hand the contour obtained from the printed pattern, and on the other hand the intended pattern and/or the simulated contour. A direct comparison can be made between simulation and experiment, without losing all the pieces of info contained in each single CD-SEM picture.