Substrate Film Thickness Estimation With Color-Change Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for calculating film thickness on a substrate fail to accurately consider the influence of underlying films, leading to inaccurate thickness estimation due to variations in the base portion's state.
Innovation Solution
A method involving a color-change estimation model to estimate color variations on the substrate surface post-processing, followed by a correlation estimation model to correlate these changes with the film's shape characteristic values, allowing for precise film thickness estimation without relying on pre-existing models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If film thickness is calculated using only post-processing image data, then the calculation process is simple, but the estimation accuracy is poor due to ignoring underlying film influences
Solution Approach 1:
The patent segments the film thickness estimation problem into two independent models: a color-change estimation model that handles substrate color variations, and a correlation estimation model that handles film thickness-correlation. This segmentation allows each model to focus on specific aspects, improving overall accuracy while keeping individual model complexity manageable
Solution Approach 2:
The patent performs preliminary action by creating a color-change estimation model first to predict what the post-processing image would look like based on the pre-processing image. This preliminary estimation is then used to correct the actual post-processing image before film thickness calculation, effectively removing the influence of underlying film variations before the main measurement
2Measurement precision
If pre-processing images are acquired and processed, then the influence of underlying films is considered, but the processing time and complexity increase
Solution Approach 1:
The color-change estimation model is created in advance as a preliminary step. Once created, this model can be reused for multiple post-processing images, reducing the time required for each individual film thickness measurement while maintaining high accuracy
Solution Approach 2:
The patent creates a predicted post-processing image (a copy or simulation) based on the pre-processing image and the color-change estimation model. This predicted image serves as a reference that can be compared with actual post-processing images, enabling accurate film thickness estimation without requiring complex real-time processing of underlying film variations
Data Source
AI summary
A device for creating a shape characteristic value estimation model for estimating a shape characteristic value includes: a post-processing-image acquisition part for acquiring a post-processing image of a surface of a substrate subjected to film processing; a pre-processing-image acquisition part for acquiring a pre-processing image of the surface not subjected to the film processing; a color-change estimation model creation part for creating a color-change estimation model for estimating first color-related information of the surface of the substrate included in the post-processing image from second color-related information of the surface of the substrate included in the pre-processing image; and a correlation estimation model creation part for creating a correlation estimation model by obtaining a difference between the first information and a result estimated by the color-change estimation model, the correlation estimation model being used for estimating correlation between the difference and the shape characteristic value.