Semiconductor Wafer Pattern Height Estimation via Shadow Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques for measuring the height of patterns on semiconductor wafers, such as the shadow effect method, are limited in accuracy and require precise on-wafer measurements, which is not feasible for patterns with edge widths smaller than the primary electron beam diameter.
Innovation Solution
A computer-implemented method using a shadow model to predict the height of a pattern on a semiconductor wafer by comparing a measured image with a set of predicted images generated based on estimated heights, allowing for accurate estimation without requiring precise on-wafer measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the shadow effect method is used to measure pattern height, then measurement capability is provided, but measurement precision deteriorates for patterns with edge widths smaller than the primary electron beam diameter
Solution Approach 1:
The patent creates a simulated shadow image through computational modeling that replicates the characteristics of a real shadow effect image. This simulated image serves as a reference template that can be compared against actual measured images, enabling precise height determination without directly measuring the problematic small features. The copying principle allows the system to work with idealized models rather than directly confronting measurement limitations.
Solution Approach 2:
The patent replaces the direct physical measurement approach (mechanical shadow effect method) with a computational image processing approach. Instead of relying on the physical shadow effect alone for measurement, the system uses simulated images generated through computational models and compares them with actual images using image processing algorithms. This substitution transitions from a purely physical measurement method to a hybrid computational-physical approach that overcomes the limitations of the physical method.
2Measurement precision
If precise on-wafer measurements are required for accurate height determination, then measurement accuracy improves, but device complexity and measurement feasibility worsen
Solution Approach 1:
The patent uses simulated shadow images as copies of what real measurements would produce under known conditions. These simulated images contain embedded height information that can be directly compared with actual measurements, eliminating the need for complex calibration procedures and precise on-wafer reference measurements. The copying approach allows the system to work with pre-computed reference data rather than requiring complex real-time measurement systems.
Solution Approach 2:
The patent introduces simulated shadow images as an intermediary between the actual measurement and the height determination process. Instead of directly converting raw measurement data into height values through complex calculations, the system compares actual images with simulated images that serve as a bridge. This intermediary approach simplifies the overall measurement system by providing a straightforward comparison mechanism rather than requiring complex direct measurement and calculation chains.
3Measurement precision
If the shadow effect method is used, then height measurement capability is provided, but reliability worsens due to sensitivity to measurement conditions
Solution Approach 1:
The patent performs preliminary computational work by generating simulated shadow images for a range of known heights before actual measurements are taken. These pre-computed simulated images serve as reference templates that are insensitive to measurement conditions. By preparing these references in advance, the system eliminates the need to perform complex condition-dependent calculations during actual measurement, thereby improving reliability.
Solution Approach 2:
The simulated shadow images serve as stable copies that replicate the expected measurement appearance under various known conditions. These copies are generated through computational modeling and remain consistent regardless of actual measurement variations. By comparing actual measurements against these stable simulated copies, the system achieves more reliable results that are less sensitive to variations in measurement conditions such as focus, brightness, or slight positioning errors.
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 provides a robust and reliable method for determining pattern height, independent of physical processes, with calibrations similar to other wafer inspection techniques, enabling precise measurements using SEM data without complex on-wafer measurements.
Implementation Method 1
One of the techniques to measure height is referred to as a 'shadow effect.' The technique uses a detector, located at a specific position with respect to the pattern, where the detector does not receive part of the responsive electrons because of occlusion by portions of the pattern, such as a sidewall, thereby creating a shadow.
Data Source
AI summary
A height of a pattern on a semiconductor wafer is determined by comparing a measured image of the pattern with a predicted image of the pattern, as produced by a shadow model. An estimated height of the pattern is provided as an input to the shadow model. The shadow model produces occluding contours that are used to generate predicted images. A set of predicted images are generated, each predicted image being associated with an estimated height. The estimated height corresponding to the predicted image most closely matching with the measured image is used as the height calculated by the shadow model.