Real-Time Semiconductor Pattern Inspection via CNN Image Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for simulating semiconductor circuit patterns require extensive time and resources due to the need for multiple Monte Carlo simulations across various manufacturing and image capturing processes, making real-time comparison of design and captured images impractical.
Innovation Solution
An image processing method that calculates probabilistic distribution statistics from design pattern data and process information using a CNN model, enabling the generation of estimated captured images for real-time comparison.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Monte Carlo simulation is used to simulate captured images under various manufacturing and image capturing conditions, then the accuracy of pattern matching is improved, but the calculation time increases significantly
Solution Approach 1:
The patent pre-calculates and stores captured image data under various manufacturing and image capturing conditions before actual inspection. This preliminary action allows the system to retrieve pre-computed images during inspection without performing time-consuming Monte Carlo simulations in real-time, thus maintaining high pattern matching accuracy while significantly reducing calculation time during actual use.
Solution Approach 2:
The patent creates copies of captured images under different conditions and stores them in a database. During inspection, instead of performing new simulations, the system retrieves and uses these pre-generated image copies, which dramatically reduces the time required for pattern matching while maintaining the accuracy benefits of simulated images.
2Adaptability or versatility
If multiple simulators are used to simulate different manufacturing processes (lithography, etching, image capturing), then the comprehensiveness of process parameter relationships is improved, but the total calculation time and cost increase
Solution Approach 1:
The patent merges the outputs of multiple process simulators (lithography, etching, image capturing) into a unified database of captured images. Instead of running each simulator separately and sequentially, the system integrates their results, allowing comprehensive process parameter relationship analysis to be performed more efficiently by leveraging pre-computed data from all processes simultaneously.
Solution Approach 2:
The patent creates a universal database that stores captured image data from multiple manufacturing and image capturing processes. This universal database serves multiple functions: it can be used for pattern matching, process parameter analysis, and inspection across different process stages, eliminating the need to re-run simulations for each specific application and significantly reducing total calculation time and cost.
3Ease of operation
If design pattern data is used directly as template image without considering process variations, then the simplicity of pattern matching is maintained, but the alignment accuracy deteriorates due to pattern deformation and image quality differences
Solution Approach 1:
The patent creates copies of captured images under various conditions and uses these copies as template images instead of directly using design pattern data. These template image copies incorporate actual pattern deformations and image quality characteristics from the manufacturing and capturing processes, enabling accurate alignment while maintaining the simplicity of using pre-processed images rather than raw design data.
Solution Approach 2:
The patent transforms design pattern data into template images by applying parameter changes that account for manufacturing and image capturing variations. This includes adjusting for pattern deformations, contrast changes, and other quality variations, thereby improving alignment accuracy while keeping the overall process simple by using pre-computed template images rather than complex real-time adjustments.
Data Source
AI summary
An image processing method whereby data pertaining to an estimated captured image obtained from reference data of a sample is acquired using an input acceptance unit, an estimation unit, and an output unit. The data is used when comparing the estimated image and an actual image of the sample, wherein the method includes: an input acceptance unit accepting input of the reference data, process information pertaining to the sample, and trained model data; the estimation unit using the reference data, the process information, and the model data to calculate captured image statistics representing a probabilistic distribution of values attained by the data of the captured image; and the output unit outputting the captured image statistics, and generating the estimated captured image from the captured image statistics. This permits reducing the time required for estimation and to perform comparison in real time.