Integration of Machine Learning with Semiconductor Thin Film Processing and Characterization
Faculty Mentor Information
Dr. David Estrada (Mentor), Boise State University
Presentation Date
7-2024
Abstract
The development of two-dimensional transition metal dichalcogenides for use in microelectronics has many challenges due to a lack of understanding of the relationships between processing and deposited film structure. This project aims to highlight those structure-property relationships via machine learning. The quality of grown films can be determined via Raman spectroscopy, followed by data processing modules written in Python. Using supervised learning, the Raman features will be correlated to film synthesis processes such as chemical vapor deposition (CVD) or metal-organic chemical vapor deposition (MOCVD), with the predictive output being the quality of the sample. Here, we present our preliminary results on data preparation and analysis to extract key Raman features correlated to process parameters for the implementation of robust machine learning models. Our approach aims to provide valuable information regarding the relationships between output film structure and synthesis processing parameters. to improve the efficiency of CVD and MOCVD workflows, ultimately contributing to advancements in materials science and the semiconductor industry.
Integration of Machine Learning with Semiconductor Thin Film Processing and Characterization
The development of two-dimensional transition metal dichalcogenides for use in microelectronics has many challenges due to a lack of understanding of the relationships between processing and deposited film structure. This project aims to highlight those structure-property relationships via machine learning. The quality of grown films can be determined via Raman spectroscopy, followed by data processing modules written in Python. Using supervised learning, the Raman features will be correlated to film synthesis processes such as chemical vapor deposition (CVD) or metal-organic chemical vapor deposition (MOCVD), with the predictive output being the quality of the sample. Here, we present our preliminary results on data preparation and analysis to extract key Raman features correlated to process parameters for the implementation of robust machine learning models. Our approach aims to provide valuable information regarding the relationships between output film structure and synthesis processing parameters. to improve the efficiency of CVD and MOCVD workflows, ultimately contributing to advancements in materials science and the semiconductor industry.