Etch Profile Model Optimization via Reflectance Spectra Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for optimizing plasma-assisted etch processes in semiconductor manufacturing lack accuracy and reliability, with existing etch profile models failing to provide sufficient precision for process engineers to achieve desired feature profiles.
Innovation Solution
The development of methods and systems that optimize computer models by using reflectance spectra matching and surface kinetic modeling to determine optimal etch process parameters, involving experimental measurements and computational simulations to minimize differences between experimental and computed etch profiles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing etch profile models are used, then the etch process can be simulated, but the predictive accuracy and reliability are insufficient for achieving desired feature profiles
Solution Approach 1:
The patent transforms the etch profile model from a purely geometric simulation to a physics-based model by incorporating plasma parameters (ion flux, radical flux, ion energy, radical density) as input parameters. This parameter transformation enables the model to predict etch profiles with high accuracy across different process conditions, resolving the contradiction between measurement precision and reliability.
Solution Approach 2:
The patent introduces an intermediary optimization process that uses reflectance spectra matching to bridge the gap between simulated etch profiles and actual measured profiles. By using optical measurements as an intermediary validation step, the model achieves both high predictive accuracy and reliability without requiring extensive manual tuning.
2Manufacturing precision
If process engineers manually tweak etch process parameters, then they can attempt to generate desired target feature profiles, but the process becomes difficult and time-consuming
Solution Approach 1:
The patent implements a self-optimizing model that automatically determines optimal etch process parameters by comparing simulated reflectance spectra with experimental measurements. The model performs self-validation and self-adjustment without requiring manual intervention from process engineers, thereby achieving desired etch profiles while minimizing time loss.
Solution Approach 2:
The patent incorporates a feedback mechanism where experimental reflectance spectra measurements are continuously compared with simulated spectra, and the model parameters are automatically adjusted to minimize the difference. This closed-loop feedback system enables rapid convergence to optimal parameters, eliminating time-consuming manual tweaking while maintaining manufacturing precision.
3Measurement precision
If existing etch profile models are used, then simulation can be performed, but the accuracy and reliability are not sufficient for industrial use
Solution Approach 1:
The patent applies partial optimization by focusing on the most critical plasma parameters (ion flux, radical flux, ion energy, radical density) rather than attempting to optimize all possible parameters. This selective approach maintains high predictive accuracy while avoiding excessive complexity in the optimization procedure, making the model suitable for industrial application.
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 significantly improves the predictive accuracy of etch profile models, enabling more precise control over etch processes and reducing the time and cost associated with manual parameter tuning, thereby enhancing the reliability of semiconductor manufacturing.
Implementation Method 1
receiving an experimental reflectance spectra generated from an optical measurement of an experimental etch process and also generating a computed reflectance spectra from the model
Implementation Method 2
The performance of plasma-assisted etch processes is frequently critical to the success of a semiconductor processing workflow
Data Source
AI summary
Disclosed are methods of optimizing a computer model which relates the etch profile of a feature on a semiconductor substrate to a set of independent input parameters (A), via the use of a plurality of model parameters (B). In some embodiments, the methods may include modifying one or more values of B so as to reduce a metric indicative of the differences between computed reflectance spectra generated from the model and corresponding experimental reflectance spectra with respect to one or more sets of values of A. In some embodiments, calculating the metric may include an operation of projecting the computed and corresponding experimental reflectance spectra onto a reduced-dimensional subspace and calculating the difference between the reflectance spectra as projected onto the subspace. Also disclosed are etch systems implementing such optimized computer models.