Wafer-Level Lithography Process Sequencing for Fingerprint Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lithographic processes face challenges in accurately correcting wafer-to-wafer variations and optimizing the entire lithographic process due to limited granularity of corrections and the inability to trace root causes of complex distortions, leading to suboptimal yield and efficiency in semiconductor manufacturing.
Innovation Solution
A method that involves obtaining characteristics of previous and subsequent processes, determining predicted performance parameters, and applying granular corrections on a per-wafer basis by associating measurement results with recorded context data to optimize the sequence of processes, thereby decoupling individual process contributions and minimizing wafer-to-wafer variations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If lot-level corrections are applied to wafer fingerprints, then process control is simplified, but manufacturing precision deteriorates due to inability to address wafer-to-wafer variations
Solution Approach 1:
The patent segments the correction approach from lot-level to wafer-level granularity. Each wafer receives individualized corrections based on its specific fingerprint characteristics, enabling precise addressing of wafer-to-wafer variations while maintaining manageable process control through automated per-wafer analysis and correction application.
Solution Approach 2:
The patent changes the correction parameters from generic lot-level adjustments to specific wafer-level parameters. By analyzing individual wafer fingerprints and applying tailored corrections to each wafer's distortion patterns, the system achieves higher manufacturing precision without proportionally increasing complexity.
2Manufacturing precision
If extensive measurement operations are performed to map wafer distortions, then manufacturing precision improves through accurate distortion characterization, but productivity deteriorates due to time-consuming measurement and analysis processes
Solution Approach 1:
The patent performs measurement and analysis operations in advance, before the actual lithographic exposure. By characterizing wafer distortions and determining corrections beforehand, the system ensures high measurement accuracy is achieved without compromising subsequent production throughput, as the analysis work is completed during preparatory stages.
Solution Approach 2:
The patent implements efficient measurement and analysis workflows that rapidly characterize wafer fingerprints and compute corrections. By optimizing the measurement and processing pipeline to operate at high speed, the system maintains accurate distortion characterization while minimizing the time added to each wafer's processing cycle.
3Manufacturing precision
If multivariate analysis is performed to decompose wafer data into component vectors, then manufacturing precision improves through accurate fingerprint decomposition, but loss of time increases due to computational complexity
Solution Approach 1:
The patent applies multivariate analysis selectively to extract the most significant component vectors from wafer fingerprint data. By focusing computational resources on decomposing only the dominant distortion patterns rather than attempting exhaustive analysis of all data dimensions, the system achieves sufficient decomposition accuracy while reducing computational time and resource requirements.
Data Source
AI summary
A method for optimizing a sequence of processes for manufacturing of product units, includes: associating measurement results of performance parameters (e.g., fingerprints) with the recorded process characteristics (e.g., context); obtaining a characteristic (e.g., context) of a previous process (e.g. deposition) in the sequence already performed on a product unit; obtaining a characteristic (e.g., context) of a subsequent process (e.g., exposure) in the sequence to be performed on the product unit; determining a predicted performance parameter (e.g., fingerprint) of the product unit associated with the sequence of previous and subsequent processes by using the obtained characteristics to retrieve measurement results of the performance parameters (e.g., fingerprints) corresponding to the recorded characteristics; and determining corrections to be applied to future processes (e.g. exposure, etch) in the sequence to be performed on the product unit, based on the determined predicted performance parameter.