Automated Mark Arrangement Inspection for Semiconductor Masks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In semiconductor manufacturing, the large number of alignment and measurement marks on photomasks makes visual inspection error-prone, leading to potential manufacturing delays and increased costs due to incorrect mark arrangement.
Innovation Solution
A method involving the generation of mark seed data with inspection marks that include vector information for rotation and mirroring detection, which is used to calculate and verify the correct arrangement of marks on scribe lines, employing a computer-aided approach to automate the inspection process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If visual inspection is used to check mark arrangement, then human resources can be utilized for verification, but the large number of marks (1500 or more) makes it difficult to visually find errors with high possibility
Solution Approach 1:
The patent replaces the mechanical visual inspection process with an automated computer-based inspection system. The system calculates coordinates of inspection marks from mask data, detects arrangement states by comparing with reference positions, and automatically judges correctness against arrangement check rules, eliminating the need for manual visual verification of 1500+ marks.
Solution Approach 2:
The patent transforms the inspection approach by changing from visual parameter assessment to automated coordinate-based parameter verification. By calculating precise coordinates and comparing them against predefined arrangement rules, the system converts a subjective visual task into an objective computational verification process.
2Reliability
If visual inspection is performed on 1500 or more alignment and measurement marks, then comprehensive coverage is achieved, but inspection time and human resources are significantly consumed
Solution Approach 1:
The patent replaces time-consuming manual visual inspection with automated computer processing. The system rapidly calculates coordinates, detects arrangement states, and judges correctness for all 1500+ marks automatically, reducing inspection time from potentially hours of manual work to rapid computational processing.
Solution Approach 2:
The inspection system performs self-verification by automatically comparing calculated mark arrangements against predefined arrangement check rules. The system independently identifies errors without requiring external human intervention, enabling rapid comprehensive inspection of all marks.
3Productivity
If mark arrangement errors are missed during inspection, then mask manufacturing can proceed, but defective masks are produced requiring refinement which adversely affects turn-around time and manufacturing cost
Solution Approach 1:
The patent performs preliminary automated inspection of mark arrangements before mask manufacturing proceeds. By calculating coordinates and detecting arrangement states in advance, the system identifies errors before they result in defective masks, preventing rework and maintaining manufacturing throughput.
Solution Approach 2:
The inspection system provides immediate feedback by automatically judging whether mark arrangements conform to predefined rules. This feedback mechanism enables real-time error detection and correction, preventing defective masks from being produced while maintaining manufacturing efficiency.
Data Source
AI summary
A method of inspecting a mark arrangement according to an embodiment of the present invention includes: generating mask data in which mark seed data that includes an inspection mark that includes vector information and is not drawn on a mask and mark data is arranged on a scribe line of the mask, calculating coordinates of the inspection mark from a reference position of the mark seed data, detecting an arrangement state of the inspection mark with respect to the reference position by using the coordinates and vector information, and judging whether the mark seed data is correctly arranged by comparing the arrangement state of the inspection mark with an arrangement check rule.