Machine Learning-Enabled Design of Point Defects in 2D Materials for Quantum and Neuromorphic Information Processing.

Area of Science:

• Materials Science
• Condensed Matter Physics
• Computational Chemistry

Background:

• Engineered and natural point defects in 2D materials are crucial for device properties like optoelectronics and quantum emission.
• The complexity of defects hinders experimental control and understanding of defect-property relationships at the atomic scale.

Purpose of the Study:

• To develop a rapid, predictive approach for understanding point defect properties in 2D materials.
• To identify novel defect structures for specific solid-state device applications.

Main Methods:

• Utilized deep transfer learning, machine learning, and first-principles calculations.
• Employed physics-informed featurization for a concise defect structure description.
• Analyzed defects across diverse 2D material systems.

Main Results:

• Successfully predicted key properties of point defects in 2D materials.
• Identified over one hundred promising, unexplored dopant defect structures.
• Generated a generalized understanding of defects across different material classes.

Conclusions:

• The developed approach accelerates the discovery of tailored point defects in 2D materials.
• Identified defects are suitable for quantum emission, resistive switching, and neuromorphic computing.
• This work provides a pathway for rational design of 2D materials for advanced solid-state devices.