Mask Pattern Verification Using Multi-Algorithm Feature Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mask pattern verification methods in semiconductor manufacturing are limited in detecting critical points accurately due to reliance on single feature amounts, leading to low detection accuracy and increased processing time, as they do not adequately consider the influence of adjacent patterns.
Innovation Solution
A mask pattern verification method that calculates both first and second feature amounts using multiple algorithms (CD value, NILS value, peak value, and integrated intensity value) to detect critical points, considering the influence of adjacent patterns, and allows for selection and adjustment of algorithms and thresholds for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple algorithms and feature amounts are used for critical point detection, then detection accuracy is improved, but processing time increases
Solution Approach 1:
The verification process is divided into two distinct stages: first detecting critical point candidates using a simple resist model with first feature amounts, then performing detailed verification using a detailed resist model with second feature amounts only on candidate regions. This segmentation allows the system to use multiple algorithms and feature amounts for accurate detection while limiting the time-consuming detailed verification to only necessary areas.
Solution Approach 2:
The patent applies partial action by using the detailed resist model and second feature amounts only for critical point candidate verification, not for the entire mask pattern. This partial application of the more resource-intensive verification method maintains high detection accuracy while significantly reducing overall processing time compared to applying detailed verification universally.
2Reliability
If detailed resist models are used for accurate critical point detection, then detection reliability is improved, but processing complexity increases
Solution Approach 1:
The patent segments the resist model usage into two levels: a simple resist model for initial critical point candidate detection, and a detailed resist model for verification of candidates only. This segmentation reduces the overall system complexity by limiting the detailed model's application scope while maintaining detection reliability through its use in the verification stage.
Solution Approach 2:
The first feature amounts and critical point candidate detection act as an intermediary step between the simple resist model and the detailed resist model. This intermediary layer filters the verification process, allowing the detailed model to be used effectively for reliability without being applied to the entire pattern, thus managing complexity.
3Productivity
If single feature amount verification is used, then processing speed is maintained, but detection accuracy deteriorates
Solution Approach 1:
The verification process is segmented into two phases: rapid initial screening using first feature amounts with a simple resist model to identify critical point candidates, followed by accurate verification using second feature amounts with a detailed resist model on candidates only. This segmentation enables the system to maintain high processing speed through the first phase while achieving high detection accuracy in the second phase.
Solution Approach 2:
The patent applies the more accurate but slower detailed verification method partially only to critical point candidate regions rather than the entire mask pattern. This partial application maintains overall processing speed by limiting the slower operation to small candidate regions while still achieving high detection accuracy where it matters most.
Data Source
AI summary
According to one embodiment, a mask pattern verification method includes: calculating mask pattern data; calculating an optical image and a resist image; calculating a first feature amount and a second feature amount, using a plurality of algorithms; in each of the plurality of algorithms, comparing the first feature amount with a first threshold, and detecting a critical point candidate in a first pattern; in each of the plurality of algorithms, comparing the second feature amount with a second threshold, and detecting a critical point in the first pattern; and selecting at least one of the plurality of algorithms, and displaying a detection result of the critical point corresponding to a selected algorithm.