3D Feature Profiling by SEM Cross-Section Imaging After Milling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing semiconductor fabrication inspection tools lack the imaging depth and signal extraction capability needed for effective 3D profiling of high aspect ratio features, such as those found in 3D NAND and DRAM samples, requiring high resolution, high throughput, and fast profiling methods.
Innovation Solution
A method involving the acquisition of two high-resolution SEM images of a sample surface, where the first image is taken before and the second after milling to expose cross-sections at multiple depths, allowing for the construction of a 3D model by comparing feature cross-sections across the images, utilizing charged particle tools like ion beams for delayering and glancing angle milling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If optical critical dimension (OCD) system or critical dimension scanning electron microscope (CD-SEM) is used for inspection, then measurement capability is provided, but imaging depth and signal extraction capability are insufficient for 3D profiling of high aspect ratio features
Solution Approach 1:
The patent transitions from 2D surface imaging to 3D volumetric profiling by introducing the depth dimension through sequential milling at different depths and reconstructing feature cross-sections across multiple depth levels, enabling complete 3D characterization of high aspect ratio features
Solution Approach 2:
The patent divides the high aspect ratio features into multiple cross-sectional slices at different depths, imaging each slice separately through sequential milling and then reconstructing the complete 3D structure by combining information from all slices
2Measurement precision
If existing fabrication inspection tools are used, then some measurement is possible, but throughput and analysis speed are insufficient for high resolution profiling
Solution Approach 1:
The patent performs preliminary milling to expose cross-sections at predetermined depths before imaging, allowing efficient capture of multiple depth levels in sequence rather than requiring post-imaging analysis or repeated measurements, thereby improving throughput while maintaining high resolution
Solution Approach 2:
The patent implements a continuous workflow where milling and imaging operations are performed in sequential batches across multiple depth levels without interrupting the overall measurement process, maintaining high throughput while achieving complete 3D profiling
3Loss of information
If high aspect ratio features are analyzed with existing tools, then some data is obtained, but detailed 3D property extraction is insufficient
Solution Approach 1:
The patent extracts detailed feature property information by isolating and imaging cross-sections at multiple depth levels separately, then reconstructing the complete 3D feature properties including position shifts and radius changes as functions of depth, thereby extracting comprehensive information that cannot be obtained from single-depth imaging
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 enables quick analysis of feature properties, including position shifts and radius changes over sample depth, providing detailed 3D models with reduced measurement time and accurate representation of feature properties, addressing the limitations of existing methods.
Implementation Method 1
milling the sample to remove at least a part of the first surface and expose a second surface
Implementation Method 2
milling the sample to remove at least a part of the first surface
Implementation Method 3
acquiring a first image of a first surface of the sample; acquiring a second image of the second surface
Data Source
AI summary
Multiple features in a sample are analyzed based on a first sample image of a first surface and a second sample image of a second surface. The first surface includes cross-sections of the multiple features, and the second surface includes cross-sections of the multiple features at different sample depths relative to the first surface. The second surface formed by milling the sample to remove at least a part of the first surface. A 3D model of the multiple features is constructed by comparing the cross-sections of multiple features in the second image and corresponding cross-sections of the multiple features in the first image.