Lithography Predictive Maintenance Using Transfer Entropy Causality

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Lithographic systems face challenges in predictive maintenance due to difficulties in accurately determining context transitions, identifying causal relationships between parameters, and managing unattended alerts, leading to inefficiencies in process control and fault diagnosis.

Innovation Solution

The method involves determining transfer entropy between pairs of time series to identify causal relationships, applying quality weightings to context data based on accuracy, and managing alerts by evaluating cost and benefit metrics to prioritize maintenance actions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional monitoring methods are used to track system parameters, then the system can operate continuously, but the accuracy of context determination and causal relationship identification deteriorates

Engineering Contradiction:

Improvecontext determination accuracyVSAvoidcausal relationship identification

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The patent introduces transfer entropy as an intermediary metric to analyze relationships between system parameters. This mathematical tool acts as a mediator that quantifies the directional information flow between parameters, enabling accurate identification of causal relationships without direct intervention in the system operation.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces traditional mechanical monitoring approaches with information-theoretic methods. Instead of using physical sensors and direct measurement systems, the invention uses transfer entropy calculations to detect causal relationships, substituting mechanical monitoring with computational analysis of parameter time series.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If comprehensive monitoring of all system parameters is implemented, then fault diagnosis capability is improved, but system complexity and computational requirements increase

Engineering Contradiction:

Improvefault diagnosis capabilityVSAvoidmonitoring system complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies local quality by focusing monitoring efforts on specific critical relationships between parameters rather than uniformly monitoring all parameters. Transfer entropy calculations identify which parameter pairs have significant causal relationships, allowing the system to concentrate analytical resources on locally important interactions rather than globally analyzing all parameters equally.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent segments the monitoring task by dividing it into individual transfer entropy calculations for different parameter pairs. This segmentation allows the complex problem of comprehensive monitoring to be broken down into manageable computational units, where each parameter pair can be analyzed independently through transfer entropy metrics.

Inventive Principle:

Principle #1Segmentation

3Measurement precision

If continuous monitoring and analysis of all alerts is performed, then predictive maintenance accuracy is improved, but processing time and computational resources increase

Engineering Contradiction:

Improvepredictive maintenance accuracyVSAvoidalert processing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies partial action by analyzing only the most relevant parameter relationships using transfer entropy, rather than performing exhaustive analysis of all possible parameter combinations. This selective approach maintains high predictive accuracy by focusing on critical causal relationships while reducing overall computational burden and processing time.

Inventive Principle:

Principle #16Partial or excessive action

4Measurement precision

If multiple context segments are analyzed with quality weighting, then measurement accuracy is improved, but computational complexity increases

Engineering Contradiction:

Improvecontext data accuracyVSAvoidcomputational complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent changes the parameter of context segmentation by dividing the operational context into multiple segments and assigning quality weights to each. This parameter transformation allows the system to account for varying data quality across different operational phases, improving overall measurement accuracy through weighted aggregation of segment-specific analyses.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS11543814B2Methods of modelling systems or performing predictive maintenance of lithographic systems

Publication Date: 2023.01.03 ASML NETHERLANDS BV

AI summary

Predictive maintenance methods and systems, including a method of applying transfer entropy techniques to find a causal link between parameters; a method of applying quality weighting to context data based on a priori knowledge of the accuracy of the context data; a method of detecting a maintenance action from parameter data by detecting a step and a process capability improvement; a method of managing unattended alerts by considering cost/benefit of attending to one or more alerts over time and assigning alert expiry time and/or ranking the alerts accordingly; a method of displaying components of a complex system in a functional way enabling improvements in system diagnostics; a method of determining the time of an event indicator in time series parameter data; a method of classifying an event associated with a fault condition occurring within a system; and a method of determining whether an event recorded in parameter data is attributable to an external factor.