Reticle Inspection Thin Line Detection Using Band-Limited Spot Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional reticle inspection methods face challenges in distinguishing between printable and non-printable features, particularly thin lines, due to optical limitations and complexities introduced by Resolution Enhancement Techniques like OPC, leading to false positives and inadequate segmentation.
Innovation Solution
The method involves generating a band-limited spot image from transmitted and reflected optical images, calibrating to minimize optical aberrations, and restoring the image to improve segmentation and line width measurement, using level-set functions to differentiate between thin and non-thin lines, and preventing thin line growth from encroaching on non-thin features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional inspection methods use user defined characteristics for differentiating features, then the inspection process is simple, but the segmentation between printable and non-printable features is inadequate leading to false positives
Solution Approach 1:
The patent performs preliminary image processing operations (optical aberration calibration, image restoration) before the actual inspection to improve the quality of input images. This preliminary preparation enables more accurate segmentation and line width measurement, resolving the contradiction by enhancing measurement precision through pre-processing rather than during the main inspection process.
Solution Approach 2:
The patent introduces processed images as an intermediary between the raw optical images and the final inspection results. By calibrating optical aberrations and restoring images to improve segmentation, the processed images serve as a mediator that enhances the accuracy of thin line detection while managing complexity through systematic image processing steps.
2Measurement precision
If optical images are used directly for thin line detection, then the process is fast, but the line width measurements do not correlate well with the underlying mask pattern
Solution Approach 1:
The patent performs optical aberration calibration and image restoration as preliminary actions before measurement. This pre-processing ensures that subsequent line width measurements accurately reflect the underlying mask pattern, resolving the contradiction by investing time upfront to gain measurement precision without repeating processing for each measurement.
Solution Approach 2:
The patent replaces direct mechanical/optical measurement with a processed image-based measurement system. By calibrating optical aberrations and restoring images computationally, the system achieves accurate line width measurements that correlate with the mask pattern, substituting computational processing for direct optical measurement limitations.
3Reliability
If thin line detection is performed without preventing thin line growth, then detection sensitivity is high, but large geometries are encroached upon causing false positives
Solution Approach 1:
The patent applies different processing and protection strategies to different regions of the image. By identifying and protecting large geometries specifically while maintaining thin line detection sensitivity in other areas, the system achieves reliable detection without false positives. This local differentiation resolves the contradiction by allowing high sensitivity where needed while protecting vulnerable regions.
4Manufacturing precision
If OPC modifications are applied to mask patterns, then optical limitations are overcome, but the mask patterns become more complex and dissimilar to resulting wafer images
Solution Approach 1:
The patent introduces processed images as an intermediary that bridges the gap between complex OPC-modified mask patterns and the resulting wafer images. By calibrating optical aberrations and restoring images, the system creates a representation that maintains the accuracy benefits of OPC while reducing the complexity and dissimilarity issues, enabling reliable inspection despite OPC modifications.
Data Source
AI summary
A detection method for a spot image based thin line detection is disclosed. The method includes a step for generating a band limited spot image from a transmitted and reflected optical image of the mask. The spot image is calibrated to minimize a plurality of optical aberrations from the spot image. The spot image is restored back to a mask image to allow at least one of: a more reliable segmentation between thin line and non-thin line areas on the mask image or a more accurate line width measurement for facilitating segmentation. Thin line features and non-thin lines features are distinguished on the restored mask image. Areas containing thin line features are grown while preventing the thin line growth from encroaching the non-thin line features.