Semiconductor Wafer Inspection via Slope Data Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automated optical inspection methods for semiconductor substrates are inadequate for reliably detecting hairline cracks and edge fractures due to the complexity of wafer surfaces and undefined edge exclusion areas, often requiring manual inspection.
Innovation Solution
An automated inspection method using an inspection apparatus that records slope data along specific directions around the center point of the semiconductor substrate, processing this data to identify locations with predetermined criteria, and utilizing a deflectometry apparatus with a light emitting line to detect hairline cracks by analyzing slope changes orthogonal to the cracks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated optical inspection methods are used, then productivity increases, but detection reliability deteriorates due to complex wafer surfaces and edge exclusion areas
Solution Approach 1:
The inspection area is segmented into a plurality of partial inspection areas arranged radially around the center point, with each partial inspection area covering a specific angular range. This segmentation allows the system to systematically inspect the entire annular region while avoiding the problematic edge exclusion area, thereby maintaining both high productivity through automation and reliable defect detection.
Solution Approach 2:
The inspection system applies local quality by using direction-specific evaluation vectors that are adapted to the local geometry of each measuring point. At each point, the evaluation vector is oriented based on the local surface normal and radial direction, allowing the system to optimize detection sensitivity for hairline cracks in different orientations while maintaining consistent performance across the entire inspection area.
2Reliability
If manual inspection is used, then detection reliability improves, but productivity deteriorates due to labor-intensive processes
Solution Approach 1:
The system replaces manual mechanical inspection with an automated optical measurement system that uses non-contact optical sensors to detect surface defects. The automated evaluation algorithm processes measurement data from multiple partial inspection areas, replicating and enhancing the capabilities of manual inspection while dramatically increasing throughput and eliminating human fatigue and variability.
Solution Approach 2:
The inspection system changes parameters by evaluating surface measurements in multiple directions (radial and tangential components) rather than relying on a single inspection direction. This multi-directional parameter evaluation enables the automated system to detect hairline cracks regardless of their orientation, matching the adaptability of manual inspection while maintaining automated speed.
3Area of stationary object
If inspection area includes edge exclusion area, then measurement coverage improves, but detection precision deteriorates due to undefined edge characteristics
Solution Approach 1:
The system extracts and excludes the problematic edge exclusion area from the inspection region by defining an annular inspection area that stops before reaching the wafer edge. This allows the system to focus measurements on the reliable annular region where surface characteristics are well-defined, thereby maintaining high measurement precision while still providing comprehensive coverage of the usable wafer area.
Solution Approach 2:
The inspection system transitions from a single radial inspection approach to a multi-dimensional approach by dividing the inspection area into multiple partial inspection areas with different angular ranges. Each partial area is inspected with locally optimized evaluation vectors, creating a comprehensive three-dimensional measurement coverage that improves precision without requiring inspection of the problematic edge region.
4Measurement precision
If slope data is evaluated in multiple directions, then detection sensitivity improves, but device complexity increases
Solution Approach 1:
The data processing is segmented by dividing the inspection area into multiple partial inspection areas, with each area processed independently using pre-defined evaluation vectors. This segmentation allows the complex multi-directional slope analysis to be broken down into manageable computational tasks that can be executed efficiently, reducing overall processing complexity while maintaining high detection sensitivity.
Solution Approach 2:
The system performs preliminary action by pre-calculating and storing optimal evaluation vector orientations for each measuring point based on the known wafer geometry and expected crack orientations. During actual inspection, the system only needs to evaluate slope data along these pre-determined directions, significantly reducing computational complexity while maintaining high detection sensitivity for hairline cracks.
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
Enables reliable, automated detection of hairline cracks and edge fractures with high sensitivity, reducing the need for manual inspection and improving detection accuracy on patterned and rough surfaces.
Implementation Method 1
a deflectometry apparatus with a light emitting line to detect hairline cracks by analyzing slope changes orthogonal to the cracks
Data Source
AI summary
An inspection method for semiconductor substrates using slope data and corresponding inspection apparatus are provided. The inspection method includes recording, by using an inspection apparatus, first data from measuring points in an inspection area of a main surface of a semiconductor substrate. The inspection area winds around a center point of the main surface. The first data includes information about a slope of the main surface at the measuring points along a first direction that deviates from a direction tangential to a circle that contains the measuring point and that has its center in the center point by not more than ±60°. A data processing apparatus analyzes the first data to obtain position data of locations on the main surface, at which the first data fulfills predetermined criteria. The position data is output through a data interface unit.