Reticle Instance Defect Location Alignment for Wafer Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current repeater defect detection methods for semiconductor wafers struggle with accurately determining relative defect locations across different reticle instances, leading to increased false repeater detection due to large swath location accuracy discrepancies, which affects the precision of defect analysis.
Innovation Solution
A system and method that utilize an inspection subsystem with energy sources and detectors to generate and analyze output from wafers, applying computer subsystems to align and transform defect positions from swath coordinates of one reticle instance to another, thereby enhancing relative defect location accuracy and reducing false repeater counts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If repeater defect detection is performed using current swath-by-swath methods with premap and run-time alignment, then defect locations within each swath are aligned accurately, but defect locations between swaths on different reticle instances show large discrepancies (up to 2× swath location accuracy or about 300 nm)
Solution Approach 1:
The patent divides the wafer into multiple swaths and processes each swath separately through alignment and coordinate transformation. By segmenting the defect detection process into swath-level operations with individual alignment transformations, the system achieves accurate defect location determination across different reticle instances while maintaining computational efficiency.
Solution Approach 2:
The patent introduces an intermediary coordinate transformation process that maps defect locations from swath coordinates to reticle instance coordinates. This intermediary transformation layer reconciles the alignment discrepancies between swaths and enables accurate repeater detection by providing a common reference frame for comparing defect locations across different reticle instances.
2Reliability
If a large repeater tolerance (e.g., 300 nm or 10 pixels) is used to account for location discrepancies, then all repeater instances can be detected, but more random defects are falsely detected as repeaters
Solution Approach 1:
The patent implements a feedback mechanism where defect locations are transformed to reticle instance coordinates and compared against a database of previously detected defects. The system uses this feedback to identify true repeaters by matching transformed coordinates with existing defect records, thereby reducing false positives while maintaining high detection completeness.
Solution Approach 2:
The patent creates a copied reference frame by transforming all defect locations to a common reticle instance coordinate system. This copied coordinate system serves as a standardized reference that enables accurate comparison and matching of defect locations across multiple reticle instances, eliminating the need for large tolerance values.
3Ease of manufacture
If defect locations are reported in swath coordinates without transformation, then the reporting process is simple, but the relative defect locations between different reticle instances lack accuracy
Solution Approach 1:
The patent performs preliminary coordinate transformation operations during the defect detection process itself, converting swath coordinates to reticle instance coordinates before defect reporting. This preliminary action ensures that accurate relative defect locations are established early in the process, eliminating the need for subsequent complex post-processing transformations.
Solution Approach 2:
The patent replaces the mechanical coordinate reporting system with a transformed coordinate system that automatically accounts for alignment variations. By substituting the simple but inaccurate swath coordinate reporting with a transformed reticle instance coordinate system, the patent achieves both accuracy and computational efficiency through algorithmic coordinate transformation.
Data Source
AI summary
Methods and systems for transforming positions of defects detected on a wafer are provided. One method includes aligning output of an inspection subsystem for a first frame in a first swath in a first die in a first instance of a multi-die reticle printed on the wafer to the output for corresponding frames, swaths, and dies in other reticle instances printed on the wafer. The method also includes determining different swath coordinate offsets for each of the frames, respectively, in the other reticle instances based on the swath coordinates of the output for the frames and the corresponding frames aligned thereto and applying one of the different swath coordinate offsets to the swath coordinates reported for the defects based on the other reticle instances in which they are detected thereby transforming the swath coordinates for the defects from swath coordinates in the other reticle instances to the first reticle instance.