Using Machine Learning to Extract Silicon Bulk Defect Parameters: A Case Study Using the Shockley-Read-Hall Equation
Description
To reduce the cost of silicon solar cells and improve their efficiency, it is crucial to identify and understand the defects limiting the electrical performance in silicon wafers. Bulk defects in semiconductors produce discrete energy levels within the bandgap and may act as recombination centers. This project investigates the viability of using machine learning for characterizing bulk defects in Silicon by using a Random Forest Regressor to extract the defect energy level and capture cross section ratios for a simulated Molybdenum defect and experimental Silicon Vacancy defect. Additionally, a dual convolutional neural network is used to classify the defect energy level in the upper or lower half bandgap.
Date Created
The date the item was original created (prior to any relationship with the ASU Digital Repositories.)
2023-05
Agent
- Author (aut): Woo, Vanessa
- Thesis director: Bertoni, Mariana
- Committee member: Rolston, Nicholas
- Contributor (ctb): Barrett, The Honors College
- Contributor (ctb): Electrical Engineering Program