3D Imaging for Semiconductor Wafer Defect Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current semiconductor inspection techniques, particularly for 3D NAND wafers, face challenges in detecting defects buried within thick layers due to limitations in optical inspection methods, such as confocal imaging and electron beam inspection, which struggle with signal noise and throughput, and often require destructive processes or expensive electronic tests.
Innovation Solution
The implementation of a three-dimensional imaging system that captures signal propagation within thick semiconductor structures, allowing for accurate defect detection and classification by generating and processing 3D images of wafer volumes, differentiating defects of interest from nuisance signals, and determining defect locations with high precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If confocal optical inspection is employed at different layer depths, then spurious optical signals from structures above and below the focal plane are eliminated, but important signal information resulting from the interaction of structures within the focal plane and structures above and below the focal plane is also lost
Solution Approach 1:
The patent extracts and removes only the spurious optical signals from structures above and below the focal plane while preserving the important signal information. This is achieved through advanced signal processing techniques that differentiate between nuisance signals and useful signal information, allowing the system to eliminate harmful factors without losing critical data about defect depth and structure interactions.
2Productivity
If traditional optical inspection techniques are used, then inspection speed is maintained, but defects buried in relatively thick layers cannot be detected
Solution Approach 1:
The patent transitions from traditional two-dimensional optical inspection to three-dimensional optical inspection by introducing the depth dimension (z-axis) through focus offset variation. This allows the system to detect defects at different depths within thick layered structures while maintaining high inspection speed through automated focus stacking and efficient data processing algorithms.
3Difficulty of detecting and measuring
If a rotating illumination beam is employed, then detection of buried defects is improved, but optical intensity is split and sensitivity is limited
Solution Approach 1:
The patent segments the illumination into multiple fixed angular directions rather than using a rotating beam. This allows simultaneous illumination from multiple angles, improving buried defect detection without splitting the optical intensity in time. The segmentation approach maintains high sensitivity while enabling comprehensive detection of defects at various depths and orientations.
4Difficulty of detecting and measuring
If electronic tests are employed to detect buried defects, then defect detection is achieved, but all device layers must be fabricated first resulting in lost time and expense
Solution Approach 1:
The patent performs optical inspection at intermediate stages during fabrication rather than waiting for complete fabrication. By using three-dimensional optical imaging to detect buried defects early in the production cycle, the system enables preliminary detection that can identify defects before additional fabrication steps are performed, reducing time and expense associated with rework or scrap.
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 effective detection and classification of defects in 3D NAND wafers with thick layered structures, improving defect detection accuracy and throughput while minimizing noise interference, and allowing for early defect identification in the production cycle without destructive processes.
Implementation Method 1
an optical subsystem configured to provide an amount of illumination light to a vertically stacked structure disposed on a substrate and collect light from the vertically stacked structure in response to the provided amount of illumination light
Data Source
AI summary
Methods and systems for improved detection and classification of defects of interest (DOI) on semiconductor wafers based on three-dimensional images are described herein. Three dimensional imaging of volumes of thick, layered structures enables accurate defect detection and estimation of defect location in three dimensions at high throughput. A series of images are acquired at a number of different wafer depths. A three dimensional image of a thick semiconductor structure is generated from the series of images. Defects are identified and classified based on an analysis of the three dimensional image of the thick semiconductor structure. In some examples, the three-dimensional image stack is visualized by contour plots or cross-sectional plots to identify a characteristic defect response. In some examples, the three-dimensional image is processed algorithmically to identify and classify defects. In another aspect, the location of a defect is estimated in three dimensions based on the three dimensional image.