Semiconductor Fingerprint Modeling for Faulty Tool Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing complexity and shrinking geometries in substrate processing for microelectronic devices require more precise manufacturing tools and early detection of faulty parts to prevent costly failures and ensure high yield semiconductor fabrication.
Innovation Solution
A method using fingerprint modeling to monitor and control process steps by calculating the sensitivity of process variables, comparing the difference between actual and predicted substrate fingerprints, and identifying faulty tools or defects without sensor input, allowing for early detection and isolation of root causes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional uniform threshold examination or sensor-based detection is used to detect faulty parts, then the detection method is simple to implement, but the precision and reliability of fault detection is insufficient
Solution Approach 1:
The patent creates a virtual fingerprint model that replicates the expected characteristics of a normally functioning substrate. This digital twin is then compared against actual substrate measurements to detect deviations. The copying principle allows fault detection without physical sensors on the substrate, achieving high precision through virtual-physical comparison rather than direct physical measurement.
Solution Approach 2:
The patent replaces traditional mechanical/sensor-based detection systems with a computational modeling approach. Instead of using physical sensors to detect faults, the system uses fingerprint modeling and comparison algorithms to identify deviations. This substitution of mechanical detection with computational analysis achieves higher precision while reducing the need for complex sensor installations.
2Reliability
If more sensors are installed to detect abnormalities in the machine, then the detection capability is improved, but the device complexity and manufacturing cost increase
Solution Approach 1:
The substrate itself provides the detection information through its inherent physical and chemical properties that create the fingerprint characteristics. The substrate's own response to processing conditions serves as the detection mechanism, eliminating the need for external sensors. This self-service approach achieves reliable fault detection while avoiding the complexity of sensor systems.
Solution Approach 2:
The patent introduces a computational fingerprint model as an intermediary between the substrate and the detection process. Rather than directly measuring substrate properties with sensors, the system uses the fingerprint model as a mediator to translate substrate characteristics into detectable signals. This intermediary approach improves reliability while reducing the need for complex sensor arrangements.
3Difficulty of detecting and measuring
If uniform threshold examination is used to detect faulty parts, then the detection process is simple, but the ability to pinpoint specific faulty process tools is insufficient
Solution Approach 1:
The patent segments the substrate fingerprint into multiple characteristic components, each corresponding to specific process parameters or tool characteristics. By analyzing which fingerprint segments deviate from the model, the system can identify which specific process tool or parameter is faulty. This segmentation transforms a single bulk measurement into multiple diagnostic indicators.
Solution Approach 2:
The patent applies local quality analysis by examining specific regions or characteristics of the fingerprint rather than treating the entire measurement uniformly. Different fingerprint characteristics are analyzed with different sensitivity thresholds and comparison methods tailored to their specific diagnostic value. This localized approach enables precise identification of which specific tool or process parameter is faulty.
Data Source
AI summary
Sensitivity calculations are provided of a process model through the rate of change of a model fingerprint with respect to process variables and defects. A fingerprint sensitivity table is generated, where process variables are associated with a set of fingerprint sensitivities. The fingerprint of incoming substrates is monitored through a production process by applying the same fingerprint method that is used in the process model. Calculations are made of the difference between the incoming substrate fingerprint and the process model predicted fingerprint. This difference fingerprint is compared against the table of fingerprint sensitivities to find the process variable most likely to be responsible for the difference. Spatial relationships between process variables and actual measurements on the substrate may be obtained. Correlation through fingerprint sensitivity improves the ability to pinpoint faulty process tools. The difference fingerprint may also identify the formation of defects on a substrate.