Spectroscopic Ellipsometry Model Optimization with Wavelength Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for quality control and testing in semiconductor manufacturing, such as spectroscopic ellipsometry, rely on modeling techniques and regression analysis but face challenges in accurately determining spectroscopic ellipsometry constants due to noise in measurement data, affecting the reliability of semiconductor properties like thickness and refractive index.
Innovation Solution
A method and apparatus that generate a simulation model for spectroscopic ellipsometry constant analysis by calculating simulation data, deriving errors between actual and measured data, and optimizing parameters to minimize the sum of squares of errors, with a weight applied to each wavelength to stabilize convergence, allowing for improved accuracy in semiconductor property determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If spectroscopic ellipsometry measurement is performed using conventional modeling techniques, then optical properties such as thickness and refractive index can be obtained, but measurement precision is reduced due to noise in the measurement data
Solution Approach 1:
The patent applies preliminary action by performing simulation calculations before actual measurement to generate reference data. The simulation model predicts spectroscopic ellipsometry constants under various conditions, creating a baseline that guides the interpretation of actual measurements and helps distinguish signal from noise.
Solution Approach 2:
The patent implements feedback through iterative optimization where simulation results are compared with actual measurement data, and the simulation model parameters are adjusted based on this comparison. This feedback loop continues until the simulation closely matches the measurement, thereby filtering noise and improving precision.
2Measurement precision
If optimization is performed to minimize sum of squares of errors, then accuracy of semiconductor property determination is improved, but convergence stability is reduced without wavelength weighting
Solution Approach 1:
The patent applies local quality by introducing wavelength-dependent weighting factors in the optimization process. Different wavelength regions are assigned different weights based on their reliability and information content, allowing the optimization to converge more stably by emphasizing reliable data regions while downweighting noisy regions.
Solution Approach 2:
The patent changes parameters by introducing wavelength weighting factors as additional optimization parameters. This modifies the error minimization function to account for varying data quality across the spectrum, enabling both high accuracy and stable convergence simultaneously.
3Reliability
If simulation model parameters are modified to fit measurement data, then reliability of optical property analysis is improved, but model complexity increases
Solution Approach 1:
The patent uses preliminary simulation models with predefined physical structures and material properties to generate initial predictions. This preliminary action establishes a physically-based framework that maintains reliability while avoiding the need for overly complex empirical models.
Solution Approach 2:
The patent creates a simplified computational copy of the physical measurement process through the simulation model. This virtual replica allows parameter optimization without requiring complex physical modifications, maintaining model simplicity while improving reliability through iterative fitting.
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 enhances the accuracy and reliability of semiconductor quality control by stabilizing convergence and reducing noise impact, enabling precise measurement of properties like thickness and refractive index, thus improving the overall quality and testing of semiconductors.
Implementation Method 1
The spectroscopic ellipsometry is an optical measurement method in which polarized light controlled by a polarizer is incident on a thin film, reflected or transmitted through the surface, and then the modulated polarization state is measured according to the refractive index or thickness of the thin film
Implementation Method 2
the modulated polarization state is measured according to the refractive index or thickness of the thin film to acquire optical properties of the thin film
Implementation Method 3
the spectroscopic ellipsometry is obtained by using a modeling technique and a regression analysis method to obtain optical properties of the thin film of the measured spectroscopic ellipsometry constant data by the change in polarization status
Data Source
AI summary
A method of quality control and testing for manufactured semiconductors by generating a model for spectroscopic ellipsometry constant analysis includes: calculating, by an operator, simulation data representing a spectroscopic ellipsometry constant of an analysis target through reasoning on attribute data representing an attribute of the analysis target using a simulation model; calculating, by an error calculator, an error representing a difference between actual measured data obtained by measuring the spectroscopic ellipsometry constant of the analysis target and the simulation data; and performing optimization in which an optimizer modifies a parameter of the simulation model according to the error.