CNN Etching Model for Nozzle Motion and Film Thickness Uniformity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The etching process in semiconductor manufacturing is complex, requiring time-consuming trial and error to optimize nozzle movement for uniform film thickness, and existing machine learning methods struggle with high-dimensional training data, limiting the effectiveness of learning models.
Innovation Solution
A training device and method using a convolutional neural network to predict film thickness changes over time, enabling the generation of learning models that optimize nozzle movement and processing conditions for uniform film thickness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of dimensions of training data is increased to represent complicated nozzle work, then the representation accuracy of processing conditions is improved, but the number of data sets required for machine learning increases exponentially
Solution Approach 1:
The patent transforms the high-dimensional time-series data of nozzle work into a lower-dimensional representation by treating time as a sequential dimension rather than expanding feature dimensions. This allows the model to capture temporal patterns without requiring exponential amounts of training data, resolving the contradiction between representation accuracy and data quantity requirements
2Manufacturing precision
If trial and error methods are used to optimize nozzle work for uniform film thickness, then the manufacturing precision is improved, but the time and cost required for optimization increases
Solution Approach 1:
The system uses machine learning models to automatically determine optimal nozzle work parameters based on training data, enabling the system to self-optimize without requiring manual trial and error by engineers. This automates the optimization process, significantly reducing both time and cost while maintaining high manufacturing precision
Solution Approach 2:
The patent implements a feedback mechanism where the actual film thickness measurement results are fed back into the machine learning model to continuously improve the optimization of nozzle work parameters. This closed-loop system enables progressive refinement of processing conditions without requiring extensive manual trial and error
Data Source
AI summary
A training device includes an experimental data acquirer that acquires a first processing amount indicating a difference between a film thickness obtained before a process for a film and a film thickness obtained after the process for the film, after a substrate processing apparatus is driven according to processing conditions including a variable condition that varies over time and executes the process for the film, and a model generator that generates a learning model, with the learning model executing machine learning using training data that includes the variable condition and the first processing amount corresponding to the processing conditions and predicting a second processing amount that indicates a difference between a film thickness obtained before the process for the film and a film thickness obtained after the process for the film wherein the learning model includes a first convolutional neural network.