QUANTITATIVE PREDICTION OF THE ELECTRICAL PROPERTIES WITH OPTICAL MEASUREMENT VIA DEEP NEURAL NETWORKS

Abstract

In the process of commercializing a material as an electrical device, countless devices are manufactured repeatedly to generalize and improve the performance of the device. Although the device fabrication is a costly and time-consuming process, there is no choice due to a lack of information about the materials. In other words, it is impossible to predict which materials are well-synthesized unless their electrical properties are measured in the form of the device. Therefore, if electrical properties could be predicted without the fabrication process, our research would be simpler, more efficient, and more reliable. For this reason, several approaches have been attempted to predict electrical properties through X-ray diffraction, atomic force microscopy, absorbance, and photoluminescence (PL). For example, the amount of doping in materials can be roughly predicted from PL peak shifts, and absorbance represents the energy and density of deep level traps of materials. These methods enable us to estimate device performance qualitatively, however, quantitative prediction of the electrical characteristics of the device is still limited due to large error deviations and poor material reliability. Therefore, the device fabrication is still inevitable to verify the electrical properties of materials, and quantitative prediction of electrical properties without the fabrication process remains a major challenge.

Document Details

Document Type
DoD Grant Award
Publication Date
Aug 11, 2021
Source ID
FA23862014085

Entities

People

  • Mun Seok Jeong

Organizations

  • Air Force Office of Scientific Research
  • Sungkyunkwan University
  • United States Air Force

Tags

Fields of Study

  • Materials science

Readers

  • Materials Science and Engineering.
  • Neural Network Machine Learning.
  • Semiconductor Device Technology

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Neural Networks