Roughness Metrology for IC Features Using Fitted Curves
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current metrics for quantifying the roughness of integrated circuit features, such as line edge roughness, are inadequate in accurately measuring the complexity of non-linear and non-circular shapes, particularly as dimensions decrease, leading to increased issues in precision and consistency.
Innovation Solution
A method involving a scanning electron microscope to image features, fitting points to a non-circular, non-linear shape, calculating perpendicular distances, and deriving metrics like standard deviation, correlation length, and Fourier analysis to quantify roughness, with optional filtering based on manufacturing processes, to provide internally consistent and effective measures of feature roughness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional roughness metrics (e.g., 3σ from straight line fitting) are used for linear edges, then measurement simplicity is maintained, but measurement precision deteriorates for non-linear and non-circular shapes
Solution Approach 1:
The patent transforms the roughness measurement from using a fixed straight line reference to using a fitted curve (defined by multiple parameters such as polynomial coefficients or control points) that adapts to the actual edge geometry. This parameter change allows the reference to match complex non-linear shapes while maintaining the statistical calculation framework (standard deviation from the fitted curve), thereby improving measurement precision without excessive complexity increase.
Solution Approach 2:
The patent segments the edge measurement into distinct phases: first fitting a curve to the edge points to capture the overall shape, then calculating deviations from this fitted curve. This segmentation allows complex shape handling to be separated from the statistical roughness calculation, improving both precision and managing complexity through modular processing.
2Productivity
If dimension reduction is applied to integrated circuit features, then circuit density is improved, but roughness effects worsen due to lack of proportional reduction
Solution Approach 1:
The patent applies local quality by allowing the fitted curve to adapt locally to the specific geometry of each edge segment rather than using a global straight line reference. This local adaptation ensures that the roughness measurement accurately reflects the actual manufacturing variations at each location, providing the precision needed to control roughness effects as dimensions are reduced for higher circuit density.
3Adaptability or versatility
If straight line fitting is used for edge roughness measurement, then ease of operation is maintained, but adaptability worsens for non-linear and non-circular features
Solution Approach 1:
The patent creates a universal measurement approach that can handle multiple feature types (linear edges, non-linear edges, circular features) through a single fitted-curve framework. The same basic process of fitting a curve and calculating deviations works across different geometries, providing adaptability while maintaining operational simplicity through a unified methodology rather than requiring different procedures for different shapes.
Data Source
AI summary
A method for evaluating a feature. The method includes receiving an image of the feature and determining respective coordinates of a plurality of points on an edge of the feature in the image. A figure having a non-circular and non-linear shape is fitted to the plurality of points, and respective distances between the plurality of points and the figure are determined. A roughness parameter for the feature is computed using the respective distances. The method may be applied in the analysis of critical dimensions (CD) of integrated circuits and, particularly, in the measurement of the edge roughness of their features and components as imaged using electron scanning microscopy (SEM).