Related Experiment Video
Updated: Aug 15, 2025
11:14
Comprehensive Characterization of Extended Defects in Semiconductor Materials by a Scanning Electron Microscope
Published on: May 28, 2016
13.9K
Learning-based physical models of room-temperature semiconductor detectors with reduced data.
Srutarshi Banerjee1, Miesher Rodrigues2, Manuel Ballester3
1Northwestern University, 2145 Sheridan Road, Evanston, IL, 60208, USA. srutarshibanerjee2022@u.northwestern.edu.
Scientific Reports
|January 4, 2023
Summary
This study introduces a novel learning-based physical model for room-temperature semiconductor radiation detectors (RTSD). It enables precise characterization of charge transport properties and material defects, even with limited data.
Area of Science:
- Materials Science
- Semiconductor Physics
- Detector Technology
Background:
- Room-temperature semiconductor radiation detectors (RTSDs) are crucial for applications in medical imaging, security, and astrophysics.
- Current characterization methods for RTSDs like CdZnTe and CdTe are labor-intensive, focusing on bulk properties rather than micron-level details.
- Sub-pixel level 3-D event reconstruction requires detailed understanding of material defects and charge transport properties.
Purpose of the Study:
- To develop a microscopic, learning-based physical model for RTSDs that overcomes limitations of current characterization techniques.
- To enable precise material property characterization using limited experimental data.
- To facilitate sub-pixel level analysis for improved 3-D event reconstruction in pixelated detectors.
Main Methods:
- A novel learning-based physical model incorporating charge transport with trapping centers was developed.
- Material properties were learned indirectly from measurable electrode signals and charge distributions during electron-hole pair injection.
- The RTSD was spatially segmented into voxels, with material properties modeled as learnable parameters within each voxel.
Main Results:
- The model successfully characterized charge drifts, trapping, detrapping, and recombination coefficients.
- Characterization was achieved considering multiple trapping centers or a single equivalent trapping center.
- The model's ability to characterize the RTSD ranged from complete to equivalent, depending on the amount of training data used.
Conclusions:
- The proposed learning-based physical model offers an efficient and data-driven approach for characterizing RTSDs at a microscopic level.
- This method significantly reduces the labor intensity associated with traditional characterization techniques.
- The findings pave the way for enhanced 3-D event reconstruction and improved performance in various RTSD applications.
