3D Wafer Volume Inspection with Averaged Cross-Section Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for inspecting three-dimensional semiconductor structures in wafers face challenges in achieving high accuracy and speed due to limitations in image acquisition time, noise ratio, and drift issues, particularly as feature sizes approach sub-nanometer scales, leading to inefficient data collection and potential damage to the wafer during the slice-and-image approach.
Innovation Solution
The method involves forming a stack of averaged image slices by fast milling and scanning of cross-section surfaces, using a moving mean value to reduce noise and compensate for drift, and optimizing image acquisition with convolution kernels or machine learning algorithms to enhance throughput and accuracy, allowing for more efficient inspection of semiconductor wafers without the need for extensive image scanning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional slice-and-image approach is used for 3D inspection, then measurement precision can be achieved, but inspection time becomes excessively long and wafer damage increases
Solution Approach 1:
The patent applies preliminary action by performing fast milling to create cross-section surfaces before detailed imaging. Multiple cross-sections are rapidly prepared in advance, allowing subsequent imaging to be faster and more efficient. This pre-preparation of sample surfaces enables the inspection process to proceed more quickly while maintaining measurement precision.
Solution Approach 2:
The patent segments the inspection process into multiple independent cross-section imaging steps. Instead of imaging the entire 3D volume in one slow process, the wafer is divided into multiple thin cross-sections that are imaged separately and then reconstructed computationally. This segmentation allows each section to be imaged faster and enables parallel processing of different sections.
2Measurement precision
If traditional slice-and-image approach is used for 3D inspection, then measurement precision can be achieved, but wafer damage increases
Solution Approach 1:
The patent performs preliminary fast milling to create the necessary cross-section surfaces before imaging, rather than slowly removing material during the inspection process. This pre-preparation minimizes the total time the wafer is exposed to damaging milling operations while still achieving the required measurement precision.
Solution Approach 2:
The patent creates multiple copies of cross-sectional views at different depths and reconstructs the 3D structure computationally from these 2D copies. This approach allows accurate 3D measurement without requiring physical slicing or extensive material removal, thereby reducing wafer damage while maintaining measurement precision.
3Productivity
If fast milling and scanning are used to reduce acquisition time, then productivity increases, but noise ratio worsens
Solution Approach 1:
The patent merges multiple fast-acquired cross-section images into a single reconstructed 3D model. By combining information from multiple noisy 2D slices, the computational reconstruction produces a higher quality 3D representation with improved signal-to-noise ratio. The merging process allows fast acquisition while recovering information quality through data integration.
Solution Approach 2:
The patent creates multiple copies of cross-sectional data at different depths and uses computational algorithms to reconstruct the 3D structure. This copying and reconstruction approach allows fast scanning with reduced noise in individual images, while the combined 3D reconstruction recovers the lost signal quality through mathematical processing of multiple copies.
4Productivity
If fast milling and scanning are used to reduce acquisition time, then productivity increases, but drift issues worsen
Solution Approach 1:
The patent segments the inspection into multiple discrete cross-section measurements taken at different depths. Each cross-section is measured independently and quickly, minimizing the time for each individual measurement and reducing drift accumulation. The segmented approach allows drift to be corrected between sections rather than compounding over a long continuous measurement.
Solution Approach 2:
The patent performs preliminary fast milling to create all necessary cross-section surfaces before imaging begins. This pre-preparation establishes stable reference surfaces for all subsequent measurements, reducing drift issues during the imaging phase. By completing the milling beforehand, the imaging process can proceed quickly without interruption, minimizing opportunities for drift to affect measurement consistency.
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 reduces image acquisition time, improves signal-to-noise ratio, and increases the robustness of inspections, enabling more precise and efficient analysis of semiconductor structures with reduced wafer damage, thus addressing the limitations of existing methods.
Implementation Method 1
A charged particle beam is used for imaging the cross-section surfaces, and at least one detector is sensitive to secondary particles and backscattered particles
Implementation Method 2
A charged particle beam is used for imaging the cross-section surfaces, and at least one detector is sensitive to secondary particles and backscattered particles
Implementation Method 3
A focused ion beam is used for milling cross-section surfaces into the wafer
Implementation Method 4
A focused ion beam is used for milling cross-section surfaces into the wafer
Data Source
AI summary
A system and a method for volume inspection of semiconductor wafers are configured for milling and fast image acquisition of cross-sections surfaces in an inspection volume. High quality images can be obtained by restriction of the imaging to regions of interest or by averaging over several fast image scans. The method and device can be utilized for quantitative metrology, defect detection, process monitoring, defect review, and inspection of integrated circuits within semiconductor wafers.