Optical Critical Dimension Metrology Model Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current metrology techniques for inspecting complex 3-dimensional patterned structures face challenges such as long setup times and inaccurate measurements due to the use of simple geometrical primitives and the need for combining data from multiple sources.
Innovation Solution
The system optimizes the creation of an optical model for OCD measurements by utilizing image data from scanning tools like SEM or AFM to determine a geometrical model, which is then used to enhance the accuracy and speed of OCD measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If scatterometry or OCD techniques are used for high-speed measurement, then productivity is improved, but setup time increases significantly for complex 3-dimensional structures
Solution Approach 1:
The patent applies preliminary action by pre-determining the geometrical model parameters from image data (SEM/AFM) before performing OCD measurements. This pre-characterization of the structure's geometry allows the optical model to be prepared in advance, eliminating the need for time-consuming parameter optimization during actual measurements, thus resolving the contradiction between high measurement speed and long setup time
Solution Approach 2:
The patent introduces image data from scanning tools (SEM or AFM) as an intermediary to bridge the gap between physical structure and optical model parameters. This intermediary provides direct geometrical information that feeds into the optical model, avoiding the need for time-consuming inverse problem solving and enabling both fast measurements and reduced setup time
2Device complexity
If simple geometrical primitives are used for optical modeling, then device complexity is reduced, but measurement precision deteriorates for complex 3-dimensional structures
Solution Approach 1:
The patent applies parameter changes by transitioning from simple geometrical primitive parameters to detailed geometrical model parameters derived from actual image data. By extracting real geometrical parameters (contours, dimensions, shapes) from SEM or AFM images, the optical model accurately represents complex 3-dimensional structures while maintaining manageable complexity through systematic parameter extraction methods
3Ease of manufacture
If test structures are used for metrology, then ease of manufacture is improved, but measurement precision deteriorates as they fail to represent actual process behavior
Solution Approach 1:
The patent applies the taking out principle by extracting geometrical information directly from actual product structures rather than using separate test structures. By removing the need for test structures and directly measuring features inside the actual structure, the method achieves both manufacturing simplicity (no separate test structure fabrication) and high measurement precision (direct representation of actual process behavior)
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 reduces setup time and improves measurement accuracy for complex 3-dimensional structures by leveraging image data to create a more precise geometrical model for optical modeling, enabling real-time optimization and enhanced process control.
Implementation Method 1
Beam scanning techniques are based on scanning a given area of a sample with a focused beam of particles, collecting any kind of radiation produced by interaction between the beam and the sample (usually secondary particle emission)
Implementation Method 2
Scatterometry or OCD techniques are based on measurement of diffraction from a repetitive structure on a sample (grating), having periodicity in either one or two directions, and reconstruction of the geometrical parameters of a unit cell of a pattern through solving the inverse problem
Data Source
AI summary
A system for use in metrology of a patterned structure, the system which includes a data input utility configured to receive: first type of data related to the patterned structure, the first type of data was obtained by a first type of metrology system and comprises first type measurements and first geometrical information regarding the patterned structure; second type of data related to the patterned structure, the second type of data was obtained by a second type of metrology system and comprises second type measurement results and second geometrical information regarding the patterned structure; the second type of metrology system differs from the first type of metrology system, and a data processing and analyzing utility configured to determine values of parameters of interest based on the first type of data and the second type of data. The parameters of interest comprise (i) a first parameter of interest that is related to the first type of metrology and (ii) a second parameter of interest that is related to the second type of metrology; in which the first parameter of interest is a parameter of a first model for interpreting the first type of measurements; in which the second parameter of interest is a parameter of a second model for interpreting the second type of measurements; wherein the data processing and analyzing utility is configured to determine values of parameters of interest by applying an iterative process, wherein the iterative process comprises updating the first model based on the second type of data to provide a currently updated first model and updating the second model based on the first type of data. The first type measurements and the second type measurements are taken from measurement sites of a substrate that comprises the patterned structure; wherein the data processing and analyzing utility configured to adjust the first type measurements by using correlation curves to provide adjusted first type measurements, and use the adjusted first type measurements during a data interpretation process applied on the second type of data, in which the data processing and analyzing utility is configured to reduce a number of floating parameters of the data interpretation process to provide a reduced number of floating parameters by stabilizing measurements of non-floating parameters; wherein each of the non-floating parameters has a greater impact on the first type measurements than an impact of each of the reduced number of floating parameters.