Wafer Alignment Using At-Resolution Metrology on Product Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for alignment in semiconductor manufacturing face challenges such as limited accuracy due to sparse alignment marks, crosstalk between model parameters, and non-captured fingerprints, leading to sub-optimal corrections and poor fingerprint descriptions, especially in low-k1 lithography processes.
Innovation Solution
A method involving the direct measurement of product feature positions using high-resolution metrology systems like e-beam metrology, which captures images of individual product features and determines their positions relative to a reference point, allowing for the reconstruction of distortion maps and enabling high-order intra-field control without relying on dense post-exposure data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sparse alignment marks are used for wafer alignment, then the alignment process is simple and fast, but the alignment accuracy is limited and model parameters suffer from crosstalk
Solution Approach 1:
The patent applies self-service by using the product features themselves (which already exist on the substrate as part of the manufacturing process) as the measurement targets for alignment metrology. This eliminates the need for separate alignment marks and allows the product features to serve dual purposes: both as the final product elements and as the reference features for measuring substrate distortion and alignment accuracy.
Solution Approach 2:
The patent implements multi-functionality by enabling product features to serve multiple functions: they are both the functional elements of the semiconductor device being manufactured and the metrology targets for alignment measurement. This universal use of product features eliminates the need for dedicated alignment marks and provides dense measurement data across the entire substrate surface.
2Measurement precision
If dense post-exposure data is collected to improve alignment modeling, then the fingerprint description becomes more accurate, but the manufacturing throughput is adversely affected
Solution Approach 1:
The patent applies preliminary action by performing alignment measurements on product features before the substrate undergoes subsequent processing steps. The e-beam metrology system measures product feature positions on the substrate prior to any post-exposure processing, capturing the distortion fingerprint at an earlier stage when the substrate is still in a known state, thereby enabling accurate alignment modeling without delaying the manufacturing flow.
3Measurement precision
If conventional alignment methods using dedicated alignment marks are used, then the alignment process is straightforward, but the accuracy is limited by the sparse distribution of alignment marks
Solution Approach 1:
The patent uses the product features themselves as the measurement targets, allowing them to serve the dual purpose of being both the functional device elements and the alignment reference features. This self-service approach transforms the limited sparse alignment marks into dense product feature measurements, dramatically increasing the number of measurement points from a few dozen alignment marks to thousands of product features across the substrate.
Solution Approach 2:
The patent transitions from measuring only dedicated alignment marks (one-dimensional sparse sampling) to measuring product features across the entire substrate surface (two-dimensional dense sampling). By utilizing the spatial distribution of product features throughout the substrate, the system captures distortion information from many more locations, creating a comprehensive fingerprint map that accurately represents substrate-wide distortion patterns.
Data Source
AI summary
A method of determining a position of a product feature on a substrate, the method includes: obtaining a plurality of position measurements of one or more product features on a substrate, wherein the measurements are referenced to either a positioning system used in displacing the substrate in between measurements or a plane parallel to the surface of the substrate; and determining a distortion component of the substrate based on the position measurements.