Learning Model for Real-Time Substrate State Prediction in Etching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing substrate processing technologies face challenges in predicting substrate states at any timing during processing due to difficulties in collecting sufficient training data sets for high-accuracy learning models, necessitating frequent substrate removal and measurement, which disrupts the processing flow.
Innovation Solution
A method involving continuous measurement of reflected light spectra and structure parameters before and after processing, enabling the creation of a learning model that predicts substrate states in real-time by associating these data points, allowing for real-time control of processing conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If frequent substrate removal and measurement are performed to collect training data, then measurement precision is improved, but productivity deteriorates due to disrupted processing flow
Solution Approach 1:
The patent replaces physical substrate removal and measurement systems with an optical measurement system that uses reflected light spectra to monitor substrate processing states in real-time. The learning model analyzes reflected light characteristics to predict substrate states without mechanical intervention, thereby maintaining measurement precision while eliminating productivity disruption.
Solution Approach 2:
The patent introduces reflected light spectra as an intermediary medium to indirectly measure substrate states. Instead of directly measuring the substrate by removing it, the system measures the light reflected from the substrate surface, which contains information about the substrate state. This intermediary approach enables continuous monitoring without interrupting the processing flow.
2Manufacturing precision
If sufficient training data sets are collected for high-accuracy learning models, then manufacturing precision is improved, but loss of time increases due to frequent substrate removal
Solution Approach 1:
The patent enables continuous collection of training data by performing optical measurements throughout the substrate processing without interruption. The reflected light spectra are continuously acquired at different processing stages, allowing the learning model to be trained on comprehensive data while the substrate remains in the processing chamber, eliminating time loss associated with removal and reinsertion.
Solution Approach 2:
The patent performs preliminary optical measurements and data collection during the substrate processing itself, before the processing is complete. This allows training data to be gathered in advance for subsequent prediction tasks, eliminating the need for separate measurement sessions that would cause time loss.
3Manufacturing precision
If real-time prediction of substrate states is achieved, then manufacturing precision is improved by preventing over-etching and under-etching, but device complexity increases due to learning model implementation
Solution Approach 1:
The patent designs the optical measurement system to serve multiple functions: it acts as both a process monitoring tool during substrate processing and a data collection mechanism for training the learning model. The same reflected light measurement apparatus is used for real-time prediction and for generating training data, reducing the need for separate complex systems.
Solution Approach 2:
The system uses the substrate processing itself to generate training data through continuous optical measurement. The processing conditions and reflected light spectra are automatically recorded and paired to create training datasets without requiring separate measurement operations. The system serves itself by using its own operational data for model training.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables precise, real-time adjustment of substrate processing parameters, reducing defects and improving yield by preventing over-etching and under-etching, and enhancing manufacturing efficiency without interrupting the process.
Implementation Method 1
a collection system that collects reflected light reflected from an illumination region on a substrate
Implementation Method 2
a spectroscopic device that measures a reflected light spectrum
Data Source
AI summary
A method includes acquiring structure parameters and reflected light spectra before and after a change in a state of a substrate caused by substrate processing, calculating a structure parameter and a reflected light spectrum in a change period based on the acquired structure parameters and reflected light spectra before and after changing the state of the substrate, creating a learning model that receives a reflected light spectrum as an input and outputs a predicted value of a structure parameter, by performing machine learning using a training data set that includes the structure parameters and the reflected light spectra before and after changing the state of the substrate and the calculated structure parameter and reflected light spectrum and using the predicted value of the structure parameter to automatically adjust an operational parameter of a substrate processing.