Wafer Load Grid Prediction for Overlay Error Maintenance Timing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing lithographic processes face challenges in accurately predicting the impact of maintenance actions, such as wafer table replacements, on overlay errors due to the inability to isolate and quantify the contribution of wafer load grid (WLG) non-correctable errors, leading to yield loss and production delays.
Innovation Solution
A method is developed to predict the on-product parameter of interest (OPO) by isolating the contribution of wafer load grid (WLG) non-correctable errors (NCEs) through a combination of online and offline phases, involving data processing and scaling factors to determine optimal maintenance times, thereby minimizing yield loss and production downtime.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If wafer table is replaced to maintain manufacturing precision, then overlay errors are reduced, but production downtime increases due to ramp-down and ramp-up periods
Solution Approach 1:
The system performs preliminary monitoring of wafer load grid non-correctable errors during normal operation and predicts when maintenance should be performed. By using performance indicators and scaling factors to forecast error levels, the system enables proactive scheduling of wafer table replacements at optimal moments, avoiding both premature maintenance that causes unnecessary downtime and delayed maintenance that degrades manufacturing precision
Solution Approach 2:
The system continuously monitors inline on-product data and feedsvack the measured performance indicators to the prediction model. This feedback loop allows the system to track the actual wear pattern of the wafer load grid, update predictions of non-correctable errors, and dynamically determine the optimal maintenance timing based on real-time conditions rather than fixed schedules
2Manufacturing precision
If wafer table is replaced frequently to maintain overlay precision, then manufacturing precision is improved, but productivity decreases due to frequent maintenance interruptions
Solution Approach 1:
The system calculates and stores scaling factors that relate performance indicators to actual overlay errors in advance. These pre-determined scaling factors enable rapid prediction of when overlay errors will exceed thresholds, allowing maintenance to be scheduled just-in-time rather than on fixed intervals, thus maximizing productivity while maintaining precision
Solution Approach 2:
The system transitions from fixed-schedule maintenance to condition-based maintenance by continuously monitoring performance indicators and predicting overlay error trends. By changing the maintenance decision parameter from time-based to error-level-based, the system optimizes the balance between manufacturing precision and productivity
Data Source
Figure 1
Figure 2~3(b)
Figure 4
AI summary
Disclosed is a method of predicting an on-product parameter of interest data component attributable to physical interaction between a substrate and a substrate support. The method comprises obtaining inline on-product data relating to a measurement of said substrate; removing non-relevant components of said inline on-product data from said inline on-product data to obtain processed inline on-product data, said non-relevant components comprising components of said inline on-product data not attributable to the wafer load grid and/or which are correctable; determining at least one performance indicator describing a variation of the processed inline on-product data; and applying at least one scaling factor to said at least one performance indicator to obtain said on-product parameter of interest data component; the at least one scaling factor having been determined from a characteristic onsubstrate wear pattern being representative for said physical interaction between substrate and substrate support.