Physics-Based Virtual Metrology for Film Thickness Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current semiconductor manufacturing processes face inefficiencies due to the limited availability of cost and time-efficient metrology data, leading to undetected faulty substrates, as existing systems rely on either time-consuming metrology measurements or complex statistical models requiring extensive datasets.
Innovation Solution
A method utilizing sensor data and a physics-based engine to generate training sets, which are input into a statistical virtual model to predict film thickness, allowing for real-time prediction and adjustment of process parameters to achieve desired film thickness without the need for multiple virtual models or extensive datasets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If metrology equipment is used to measure film thickness, then measurement precision is improved, but productivity deteriorates due to time and cost constraints limiting analysis to only a portion of production run
Solution Approach 1:
The patent creates a virtual copy of the metrology measurement process through a virtual model that predicts film thickness using sensor data and physics-based transformations. This virtual measurement system provides measurement precision without the time and cost constraints of physical metrology equipment, enabling full production run analysis while maintaining productivity.
Solution Approach 2:
The system performs preliminary actions by training a virtual model during setup to establish predictive relationships between sensor data and film thickness. Once trained, the model can rapidly predict thickness for all substrates in real-time during production, eliminating the need for time-consuming post-production metrology measurements and enabling proactive quality control.
2Device complexity
If a single virtual model with physics-based transformation is used, then device complexity is reduced, but measurement precision may deteriorate compared to multiple specialized models
Solution Approach 1:
The patent transforms sensor data into physics-based outputs using transformation functions that convert raw sensor readings into meaningful process parameters. This parameter transformation enables a single virtual model to capture diverse deposition conditions and maintain measurement precision across varying process conditions without requiring multiple specialized models.
Solution Approach 2:
The virtual model is designed with universal applicability to predict film thickness across different deposition conditions, substrates, and process parameters. By incorporating physics-based transformations and training on comprehensive datasets, the single model achieves multi-functionality that would otherwise require multiple specialized models, reducing complexity while maintaining precision.
3Manufacturing precision
If extensive datasets are collected for training, then manufacturing precision is improved, but loss of time and storage resources increases
Solution Approach 1:
The system performs preliminary data collection and model training during setup or maintenance periods rather than during active production. This preliminary action allows comprehensive datasets to be collected and processed in advance, establishing a trained virtual model that can then rapidly predict thickness during production without causing time delays to manufacturing precision requirements.
Data Source
AI summary
A method includes obtaining sensor data associated with a deposition process performed in a process chamber to deposit film on a surface of a substrate. The method further includes generating a plurality of physics based outputs using a transformation function and the sensor data. The method further includes mapping the physics based outputs to a training set. The method further includes training a virtual model based on the training set and the sensor data, wherein the virtual model is trained to generate predictive metrology data associated with the film.