Defect Prediction Model for Lithographic Yield Control

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Lithographic apparatuses face challenges in predicting and correcting defects during the device manufacturing process, leading to reduced yield and efficiency in producing substrates with accurate circuit patterns.

Innovation Solution

A computer-implemented method involving the training of a classification model using measured process parameters and defect indications to predict defects, allowing for real-time adjustments in the manufacturing process to prevent defects and improve yield.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If a classification model is trained using measured process parameters and defect indications to predict defects, then manufacturing precision and yield are improved, but device complexity increases

Engineering Contradiction:

Improvedefect prediction accuracyVSAvoidclassification model complexity

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The classification model is trained in advance using historical process parameter data and defect indications before actual production. This preliminary training enables the model to predict defects proactively during manufacturing, improving precision without adding real-time operational complexity.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system uses measured process parameters as inputs to create a predictive copy of defect outcomes. By analyzing patterns in historical data, the model generates predictions that replicate expert defect identification capability without requiring complex real-time inspection equipment.

Inventive Principle:

Principle #26Copying

2Productivity

If real-time process adjustments are made based on defect predictions, then productivity is improved, but measurement precision requirements increase

Engineering Contradiction:

Improvesubstrate yieldVSAvoidprocess parameter measurement accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system implements a feedback loop where defect predictions based on measured process parameters trigger real-time process adjustments. The classification model continuously receives new measurements and provides predictions that feed back to control the manufacturing process, improving productivity through adaptive control.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system adjusts process parameters in real-time based on classification model predictions. When defects are predicted, the system modifies parameters such as exposure conditions or processing settings to prevent defect occurrence, thereby improving substrate yield without requiring ultra-precise measurements.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS20210405545A1Yield estimation and control

Publication Date: 2021.12.30 ASML NETHERLANDS BV

AI summary

A defect prediction method for a device manufacturing process involving production substrates processed by a lithographic apparatus, the method including training a classification model using a training set including measured or determined values of a process parameter associated with the production substrates processed by the device manufacturing process and an indication regarding existence of defects associated with the production substrates processed in the device manufacturing process under the values of the process parameter, and producing an output from the classification model that indicates a prediction of a defect for a substrate.