Maintenance is the technology of continuously monitoring the conditions of equipment and predicting the timing of maintenance for equipment. Particularly in the field of semiconductor manufacturing, where processes are automated, various methods are being tried to minimize the economic losses and maintenance costs caused by equipment failure. A new Predictive Maintenance (PdM) technique, a new method of maintenance, is introduced in this paper to develop an algorithm for predicting the failure of wafer transfer robots in advance. The acceleration sensor data used in the experiment were obtained by installing a sensor onto the wafer transfer robot. To analyze these data, the data preprocessing and FFT process were performed. These data were divided into normal data, first error data, second error data, and third error data (failure data) in stages. By clustering the data using the K-means algorithm, the center point distribution of the clusters was analyzed, and the features of the error data and normal data were extracted. Using these features, an artificial neural network model was designed to predict the point of failure of the robot. Previous research on maintenance systems of the transfer robot used fewer than 50 error data, but 1686 error data were used in this experiment. The reliability of the model is improved by randomly selecting data from a total of 2248 data sets. In addition, it was confirmed that it was possible to classify normal data and error data with an accuracy of 97% and to predict equipment failure by applying neural network modeling.
Background: This study aims to develop an integrated optical system that can simultaneously or selectively measure the signals obtained from radioluminescence (RL), thermoluminescence (TL), and optically stimulated luminescence (OSL), which are luminescence phenomena of materials stimulated by radioactivity, heat, and light, respectively. The luminescence mechanism of various materials could be investigated using the glow curves of the luminescence materials. Materials and Methods: RL/TL/OSL integrated measuring system was equipped with a X-ray tube (50 kV, 200 μA) as an ionizing radiation source to irradiate the sample. The sample substrate was used as a heating source and was also designed to optically stimulate the sample material using various light sources, such as high luminous blue light emitting diode (LED) or laser. The system measured the luminescence intensity versus the amount of irradiation/stimulation on the sample for the purpose of measuring RL, TL and OSL sequentially or by selectively combining them. Optical filters were combined to minimize the interference of the stimulation light in the OSL signal. A long-pass filter (420 nm) was used for 470 nm LED, an ultraviolet-pass filter (260-390 nm) was used for detecting the luminescence of the sample by PM tube. Results and Discussion: The reliability of the system was evaluated using the RL/OSL characteristics of Al2O3:C and the RL/TL characteristics of LiF:Mg,Cu,Si, which were used as dosimetry materials. The RL/OSL characteristics of Al2O3:C showed relatively linear dose-response characteristics. The glow curve of LiF:Mg,Cu,Si also showed typical RL/OSL characteristics. Conclusion: The reliability of the proposed system was verified by sequentially measuring the RL characteristics of radiation as well as the TL and OSL characteristics by concurrent thermal and optical stimulations. In this study, we developed an integrated measurement system that measures the glow curves of RL/TL/OSL using universal USB-DAQs and the control program.
Abstract.In plasma etch process, the optical emission spectroscope (OES) is widely used to detect plasma etch endpoint. The OES gathers the intensity of the wavelength from the radicals in the plasma chamber. In addition, the OES is very sensitive to the external elements or a particles, which means that there are diverse noise in OES data. Therefore, it is necessary to choose meaningful data, reduce noise, and reduce the quantity of data. In this paper, a new method to detect endpoint of double layer plasma etching is proposed. This algorithm uses data from OES and utilizing principle component analysis (PCA) and local outlier factor (LOF).
Since semiconductor devices are extremely integrated, process control is much more difficult in semiconductor fabrication. The optical emission spectroscopy (OES) acquires unique wavelength intensity of particles in plasma chamber, and the endpoint can be decided by utilizing its intensity. However, the endpoint detection is difficult because not only of these tremendous amount of data, but also of extremely large amount of noise in OES data. To solve these problems, the OES data of byproducts and the data of etchants are classified by SNR. And a combination of Principal Component Analysis (PCA) and Support Vector Machine (SVM) hybrid algorithm is applied to detect endpoint. The PCA was utilized to reduce dimension of the selected data, and SVM algorithm is applied to separates status between before endpoint and after endpoint. The SVM model using SNR and PCA showed excellent performance in real-time endpoint detection in plasma etching.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.