Edge Detection in Charged Particle Beam Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern photolithography faces challenges in accurately measuring critical dimensions due to the electron beam width being comparable to or larger than the feature sizes, leading to edge cross-talk and degradation in measuring feature parameters like line width, roughness, and edge slope in prior art edge detection processes.
Innovation Solution
The method involves obtaining an image of a pattern using a charged particle beam, processing it to compute the aggregate energy of spectral components, applying an n'th order polynomial filter responsive to the main direction of edges, and using dynamic programming to detect and mitigate edge paths, thereby enhancing edge detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If prior art edge detection processes are used, then the process is simple, but measurement precision deteriorates when electron beam width is comparable to or larger than feature sizes
Solution Approach 1:
The image processing is divided into distinct stages: initial edge detection, polynomial curve fitting to model edge profiles, spectral component analysis, and iterative refinement. This segmentation allows each stage to address specific aspects of the edge detection problem, improving accuracy while keeping each processing step manageable and computationally efficient.
Solution Approach 2:
The method performs preliminary actions by first detecting initial edge positions and fitting polynomial curves to model the edge profiles before conducting spectral analysis. This preliminary modeling provides a foundation for subsequent refinement steps, enabling the system to handle complex beam-width effects without requiring complete reprocessing of the entire image at full complexity.
2Measurement precision
If the electron beam width is reduced to be narrower than feature sizes, then measurement precision improves, but the interaction volume decreases leading to signal loss
Solution Approach 1:
The invention replaces the mechanical approach of reducing beam width to improve resolution with a computational approach. Instead of physically narrowing the beam to achieve better edge definition, the system uses polynomial modeling and spectral analysis to mathematically deconvolve the beam-width effect from the measured signal, thereby maintaining signal intensity while achieving accurate measurements.
Solution Approach 2:
The method changes the approach from modifying physical beam parameters (width) to adjusting processing parameters (polynomial order, spectral component selection, iterative refinement thresholds). By varying these computational parameters, the system achieves high measurement precision without compromising signal strength, as the beam width itself remains unchanged and sufficiently intense.
3Measurement precision
If iterative path mitigation is applied to detect multiple edges, then edge detection accuracy improves, but processing time increases
Solution Approach 1:
The system performs preliminary edge detection to identify candidate edge locations and fits polynomial curves to model the expected edge profiles before conducting iterative spectral analysis. This preliminary modeling provides initial estimates that guide the subsequent iterative refinement process, reducing the number of iterations needed to converge to accurate results and thereby reducing overall processing time.
Solution Approach 2:
The method applies partial mitigation by focusing computational effort on the most significant spectral components and using polynomial fitting to capture the dominant edge characteristics. Rather than performing exhaustive analysis on all possible edge configurations, the system identifies and processes the most probable edges using the polynomial models, achieving sufficient accuracy for multiple edges without the full computational burden of complete iterative mitigation for every possible edge case.
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 effectively separates pattern and noise components, ensuring accurate detection of edges and feature parameters even when the electron beam width matches or exceeds the feature size, improving measurement precision in photolithography.
Implementation Method 1
obtaining an image of the pattern; wherein the image is generated by scanning the pattern with a charged particle beam
Data Source
AI summary
A system, a non-transitory computer readable medium and a method for detecting a parameter of a pattern, the method comprises: obtaining an image of the pattern; wherein the image is generated by scanning the pattern with a charged particle beam; processing the image to provide an edge enhanced image; wherein the processing comprises computing an aggregate energy of first n spectral components of the image, wherein n exceeds two; and further processing the edge enhanced image and determining a parameter of the pattern.