Focus Measurement Model Calibration via Wafer Offset Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current focus measurement models based on FEM wafers suffer from scanner focus position ambiguities due to process variations, leading to large systematic errors in focus and dose measurements, with accuracy limited to around 4-5 nm.
Innovation Solution
The method involves printing multiple wafers with predefined focus offsets and measuring signals across sites to quantify focus inaccuracy by comparing signals, using a model-less approach that reduces ambiguity in scanner focus positions to below 2-3 nm by averaging systematic variations across wafers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If FEM wafers with predefined focus positions are used for model calibration, then focus and dose measurement can be performed, but scanner focus position ambiguities due to process variations cause large systematic errors limiting accuracy to 4-5 nm
Solution Approach 1:
The patent divides the measurement process into two independent stages: (1) measuring process variations on product wafers to create a variation map, and (2) using this map to correct focus position measurements on FEM wafers. This segmentation allows the systematic errors to be identified and corrected separately, improving measurement accuracy beyond the original 4-5 nm limitation.
Solution Approach 2:
The patent creates a copy of the process variation characteristics by measuring product wafers and generating a variation map that replicates the systematic errors. This copied information is then applied to correct the FEM wafer measurements, eliminating the need to directly measure absolute focus positions and achieving sub-2nm accuracy.
2Reliability
If random FEM is used to reduce correlation between focus/dose characteristics and process variations, then measurement reliability improves, but model calibration complexity increases
Solution Approach 1:
The patent replaces the complex random FEM approach with a systematic correction method using product wafer measurements. Instead of randomly distributing focus positions to reduce correlation, the patent uses actual measured process variations from product wafers to correct the measurements, simplifying the calibration process while maintaining high reliability.
Data Source
AI summary
A method is provided that comprises printing FEM wafers having different predefined focus offsets and multiple corresponding sites, measuring signals from the sites, and quantifying a focus inaccuracy by comparing the measured signals from the corresponding sites across the wafers.