Front-End Wafer Geometry Metrics for Overlay Error Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional metrology tools only provide sufficient measurements for predicting overlay errors early in the fabrication process, failing to account for elastic deformation that occurs as more layers are formed on the wafer, and there are no solutions for pre-fabrication wafer distortion prediction using flatness inspection measurements of a front-end processed wafer.
Innovation Solution
A method and system using a Gapi metric, calculated from front-end processed wafer measurement data, to predict in-plane distortion and sort wafers based on a correlation with back-end yield, allowing for early identification and adjustment of front-end process tools to improve wafer flatness and reduce overlay errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional metrology tools are used to measure wafer geometry, then measurements can be obtained, but they only predict overlay errors early in fabrication and fail to account for elastic deformation occurring as more layers are formed
Solution Approach 1:
The patent measures and calculates wafer geometry metrics (shape and flatness) at the front-end processing stage, before fabrication begins. This preliminary measurement allows prediction of in-plane distortion that will occur during fabrication, enabling early identification of wafers that will develop overlay errors. The Gapi metric is calculated from front-end processed wafer data to predict distortion before elastic deformation occurs during layer formation.
Solution Approach 2:
The patent transitions from measuring only surface flatness to calculating a three-dimensional Gapi metric that incorporates both shape (long wavelength deviations) and flatness (thickness variations). This multi-dimensional approach provides a more comprehensive prediction of in-plane distortion by considering both out-of-plane and in-plane geometric characteristics of the wafer.
2Measurement precision
If high accuracy inspection tools are used to measure wafer distortions, then overlay errors can be detected, but the tools require polished surfaces and measurements can only be taken after fabrication has begun
Solution Approach 1:
The patent performs wafer geometry measurements at the front-end processing stage, immediately after the wafer is processed by the front-end tool and before fabrication begins. This timing allows the wafer surface to still be in its original state without requiring polishing or fabrication layers. The Gapi metric is calculated from these early measurements to predict distortion before it occurs during fabrication.
Solution Approach 2:
The patent creates a predictive model (Gapi metric) from front-end processed wafer measurements that replicates the distortion behavior that would occur during fabrication. This computational copy of the distortion pattern allows prediction of overlay errors without requiring physical measurement during or after fabrication, saving time and eliminating the need for polished surfaces.
3Device complexity
If no pre-fabrication wafer distortion prediction is performed, then the process is simpler, but overlay errors cannot be predicted and back-end yield is reduced
Solution Approach 1:
The patent replaces complex physical measurement systems with a computational approach. Instead of using high-accuracy inspection tools during fabrication, the system uses standard front-end processed wafer measurements combined with mathematical calculations (Gapi metric) to predict distortion. This substitution of mechanical measurement with computational prediction reduces device complexity while improving reliability through early distortion prediction.
Solution Approach 2:
The patent transforms front-end processed wafer geometry parameters (shape and flatness measurements) into a predictive Gapi metric that correlates with in-plane distortion. By changing the parameters from direct physical measurements to a calculated predictive index, the system achieves distortion prediction capability without requiring complex measurement equipment, thereby improving back-end yield while maintaining simplicity.
Data Source
AI summary
A method for processing semiconductor wafers includes providing a first semiconductor wafer processed by a front-end process tool and obtaining measurement data along a surface of the first semiconductor wafer. The method also includes calculating a Gapi value of the first semiconductor wafer based on based on the measurement data, where the Gapi value is a global metric representing a difference between a raw shape of the first semiconductor wafer and an ideal shape of the semiconductor wafer. The method also includes determining whether the Gapi value of the first semiconductor wafer is within a predetermined threshold and either tuning the front-end process tool and processing a second semiconductor wafer with the tuned front-end process tool when the Gapi value is determined to be outside of the predetermined threshold, or sorting the first semiconductor wafer for polishing when the Gapi value is determined to be within the predetermined threshold.