Combined Alignment Modeling for Multi-Layer Overlay Accuracy

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing lithographic processes face challenges in accurately aligning and measuring patterns on substrates due to distortions caused by processing steps and apparatuses, leading to errors in overlay and feature placement, particularly when different modeling strategies are used for different layers.

Innovation Solution

A method for modeling alignment data involves obtaining alignment data from multiple layers, fitting separate models to each layer, determining weightings for model parameters, and combining these models to create a combined alignment model for improved accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If different modeling strategies are used for different layers, then alignment accuracy for each layer can be optimized, but errors are introduced due to inconsistencies between models

Engineering Contradiction:
Improvealignment accuracyVSAvoidmodel consistency
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent combines multiple layer-specific alignment models into a single unified alignment model. The system integrates alignment data from multiple layers and merges the individual models to create one consistent model that maintains the accuracy benefits of layer-specific modeling while eliminating the inconsistency errors that would arise from using separate models.

Inventive Principle:
Principle #5Merging (Combining)

2Measurement precision

If separate models are fitted to each layer, then layer-specific alignment characteristics are captured, but the complexity of the alignment process increases

Engineering Contradiction:
Improvelayer-specific alignment accuracyVSAvoidmodeling process complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the alignment modeling process into distinct phases: first fitting separate models to individual layers to capture layer-specific characteristics, then combining these models into a unified model. This segmentation allows the system to benefit from layer-specific optimization while managing complexity through a structured multi-step approach rather than attempting to create one complex model all at once.

Inventive Principle:
Principle #1Segmentation

3Measurement precision

If extensive measurement operations are performed prior to exposure, then alignment accuracy is improved, but processing time increases

Engineering Contradiction:
Improvealignment accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs alignment measurements and model fitting operations before the actual exposure process. By conducting extensive measurement operations and creating the unified alignment model in advance, the system ensures high alignment accuracy is achieved beforehand, allowing the exposure process itself to proceed efficiently without time-consuming measurements during production.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4700474A1A method for modeling metrology data over a substrate area and associated apparatuses
Publication Date: 2026.02.25 ASML NETHERLANDS BV
  • EP4700474A1 patent drawingFigure 1~2
  • EP4700474A1 patent drawingFigure 3
  • EP4700474A1 patent drawingFigure 4(a)~5

AI summary

Disclosed is method for modeling alignment data for exposure of a third layer on a substrate in an exposure process. The method comprises obtaining at least first layer alignment data relating to a first layer and second layer alignment data relating to a second layer; fitting a first model to the first layer alignment data to obtain a first alignment model and fitting a second model to the second layer alignment data to obtain a second alignment model; determining, for each of a plurality of different subsets of model parameters of said first alignment model and second alignment model, a respective weighting for each of said first model and second model; and combining the first model and second model in accordance with said weighting to determine a combined alignment model.