Radiation effects in microelectronic components are a significant concern for the reliability of DoD systems that operate at high altitudes or in outer space. Typical characterization efforts focus on macroscale degradation signatures from electrical measurements at device terminals. However, a comprehensive analysis of radiation-induced physical defects is not possible based solely on terminal measurements. CFD Research Corporation and Arizona State University propose a predictive modeling effort to complement a detailed experimental approach to address this challenge. We will perform detailed physics-based modeling of the radiation response of a selected semiconductor device, and use it with the electrical characterization data to guide Transmission Electron Microscopy-based nanoscale material characterization. We will utilize the device simulation and measurement data to develop Artificial Intelligence/Machine Learning-based predictive models for quantitative correlation of the nanoscale material properties with macroscale electrical properties. In Phase I, we will perform a feasibility study based on electrical and material characterization of a simple device structure and relevant radiation effect, while using the data to develop behavioral models for the radiation effects. In Phase II, we will further develop and demonstrate the predictive model using additional device structures, material systems, and radiation effects.
