Multi-Stage Lithography Overlay Correction via Fingerprint Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current multi-layer lithographic processes face challenges in accurately aligning successive layers due to non-linear distortions, making it difficult to trace and correct issues that lead to yield loss, and conventional analysis methods are time-consuming and laborious, often detecting problems late in the process.
Innovation Solution
A method that involves receiving object data from multiple stages of processing, determining fingerprints of variation, analyzing commonality across stages, and optimizing the apparatus based on commonality results to provide real-time context-driven root cause analysis and correction advice.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional analysis methods are used to trace and correct issues in multi-layer lithographic processes, then root cause analysis can be performed, but the process is time-consuming and laborious, often detecting problems late
Solution Approach 1:
The system performs preliminary analysis by collecting and analyzing measurement data from multiple layers in advance, before final overlay correction is needed. Fingerprints are determined and analyzed across layers proactively, enabling early detection of alignment issues and predictive overlay correction, rather than waiting for problems to manifest in final product inspection.
Solution Approach 2:
The system creates simplified representations (fingerprints) of complex multi-layer alignment data that capture essential variation patterns. These fingerprint copies enable rapid analysis and comparison across layers without processing the full complexity of raw measurement data, significantly reducing analysis time while maintaining diagnostic accuracy.
2Manufacturing precision
If extensive measurement and mapping operations are performed to achieve accurate overlay within a few tens of nanometers, then alignment precision is improved, but the complexity of the lithographic apparatus increases
Solution Approach 1:
The system extracts and separates the essential alignment information from complex multi-layer measurement data by determining fingerprints that represent characteristic variation patterns. This extraction process isolates the critical alignment parameters from extraneous data, enabling accurate overlay correction without requiring the apparatus to directly handle the full complexity of all measurement operations.
Solution Approach 2:
The system transforms multi-layer spatial alignment data into a different dimensional representation through fingerprint analysis. By converting complex spatial variation patterns across multiple layers into fingerprint signatures, the system enables overlay correction in a transformed parameter space, reducing the apparent complexity of the alignment problem.
3Manufacturing precision
If advanced alignment models are used to correct non-linear distortions, then overlay accuracy is improved, but it remains difficult to trace and eliminate root causes of distortions
Solution Approach 1:
The system implements feedback by continuously analyzing fingerprints from processed layers and using this information to predict and correct overlay errors in subsequent layers. The analysis of commonality and differences in fingerprints across layers provides diagnostic feedback about the state of alignment, enabling both correction of current issues and prevention of future problems.
Solution Approach 2:
The system performs preliminary diagnostic analysis by determining and comparing fingerprints across layers before final overlay correction. This advance analysis identifies patterns of distortion and their sources, preserving diagnostic information about root causes while enabling proactive correction strategies that prevent distortion propagation to subsequent layers.
Data Source
AI summary
A method of optimizing an apparatus for multi-stage processing of product units such as wafers, the method includes: receiving object data representing one or more parameters measured across the product units and associated with different stages of processing of the product units; and determining fingerprints of variation of the object data across the product units, the fingerprints being associated with different respective stages of processing of the product units. The fingerprints may be determined by decomposing the object data into components using principal component analysis for each different respective stage; analyzing commonality of the fingerprints through the different stages to produce commonality results; and optimizing an apparatus for processing product units based on the commonality results.