Semiconductor Alignment via Ineffective Cell Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In semiconductor device appearance inspection, positional errors during imaging can lead to inaccurate defect detection, especially when the chip size exceeds the camera's visual field, causing artificial defects and reduced precision in pattern shape detection, and the current methods for alignment are complex and time-consuming.
Innovation Solution
The semiconductor device is designed with unit regions containing effective cells for main current flow and ineffective cells with distinct appearances, used as alignment patterns within each Field of View (FOV), allowing for accurate image cutting and comparison to achieve precise inspection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an alignment mark is provided for each FOV to correct imaging position errors, then measurement precision is improved, but device complexity increases and inspection time lengthens
Solution Approach 1:
The patent makes the ineffective cell serve multiple functions: it acts as both a functional element (providing electrical isolation) and an alignment pattern (providing reference for image registration). This eliminates the need for separate alignment marks, reducing device complexity while maintaining measurement precision.
Solution Approach 2:
The ineffective cell inherently provides alignment functionality without requiring additional components. The unique appearance of the ineffective cell automatically serves as a distinguishable alignment pattern, allowing the inspection system to use existing structural elements for dual purposes.
2Measurement precision
If the region where fields of view overlap is enlarged to accommodate alignment errors, then measurement precision is maintained, but the number of FOVs increases and inspection time lengthens
Solution Approach 1:
The patent replaces the mechanical approach of increasing FOV overlap (which requires physical adjustment of imaging parameters and increases inspection area) with an image processing approach. By using the unique ineffective cell as an alignment pattern, the system can accurately register images through software-based trimming, maintaining precision without expanding the physical inspection region.
3Measurement precision
If a specific unit cell is registered as an alignment pattern, then measurement precision is improved, but it becomes difficult to distinguish the alignment pattern from other unit cells
Solution Approach 1:
The patent introduces asymmetry by designing the ineffective cell with a unique appearance that differs from effective cells. This asymmetric design creates a distinguishable alignment pattern that can be easily identified among identical effective cells, solving the pattern recognition problem while maintaining alignment precision.
Solution Approach 2:
The patent applies local quality by giving the ineffective cell distinct visual characteristics (different appearance) compared to effective cells. This localized differentiation allows the alignment pattern to be uniquely identified in specific regions without affecting the uniformity of other unit cells, making pattern detection straightforward.
Data Source
AI summary
A method for manufacturing a semiconductor device according to the present invention includes a manufacturing step of forming a plurality of unit regions each having a plurality of first regions serving as effective cells in which main current flows, and a second region that has an appearance different from that of the first regions and serves as an ineffective cell in which no main current flows, and an appearance inspection step including a step of imaging the unit region to obtain a captured image, a step of cutting out an inspection image from the captured image based on a position of an alignment pattern containing the second region, and a step of comparing the inspection image with a reference image.