Plasma Processing Apparatus Wavelength Selection for Prediction Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing plasma processing technologies face challenges in predicting processing result factors due to reduced correlation after model creation, failing to consider intrinsic mechanisms or properties that affect robustness, leading to increased prediction errors.
Innovation Solution
A plasma processing apparatus and analysis method that selects wavelengths or values with significant deviation in intensity distribution to predict processing results, using a correlation between spectroscope measurement data and processing result factors, thereby enhancing prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional prediction models are created using correlation between operational data and processing result data, then initial prediction accuracy is improved, but prediction error increases over time due to reduced correlation
Solution Approach 1:
The invention changes the selection criteria for wavelengths from simply high correlation to high deviation in intensity distribution. This parameter change in the selection method allows the model to capture more robust features that maintain stability over time, resolving the contradiction between initial accuracy and long-term reliability
Solution Approach 2:
The invention introduces a feedback mechanism where prediction errors are calculated and used to iteratively improve the prediction model. By continuously evaluating prediction accuracy and adjusting the model based on errors, the system maintains both high accuracy and stability over time
2Measurement precision
If wavelengths with high correlation to processing results are selected for prediction, then prediction accuracy is improved, but robustness against disturbances is reduced
Solution Approach 1:
The invention changes the selection parameter from correlation coefficient to deviation in intensity distribution. This parameter change identifies wavelengths that are inherently more stable and less sensitive to environmental disturbances, while still maintaining their ability to predict processing results accurately
Solution Approach 2:
The invention prepares for future disturbances by selecting wavelengths with high deviation, which inherently provides a buffer against correlation degradation. This prior selection strategy cushions the model against future environmental changes and maintains robustness
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
This approach achieves a high correlation between monitoring data and processing result factors, allowing for accurate prediction of processing results and improved robustness against disturbances, reducing prediction errors.
Implementation Method 1
data of light emitted from the plasma
Implementation Method 2
spectroscope measurement data obtained by measuring light emitted from plasma during the plasma processing
Data Source
AI summary
According to the present invention, a plasma processing apparatus includes an analysis unit that obtains wavelengths of the light correlated with a plasma processing result, selects, from the obtained wavelengths, a wavelength having a first factor that represents a deviation in an intensity distribution of the light and is larger than a first predetermined value, and predicts the plasma processing result using the selected wavelength, or an analysis unit that obtains values computed using each of light intensities of a plurality of wavelengths and correlated with the plasma processing result, selects, from the obtained values, a value having a second factor that represents a deviation in a distribution of the obtained values and is larger than a second predetermined value, and predicts the plasma processing result using the selected value.