High-throughput screening and machine learning classification of van der Waals dielectrics for 2D nanoelectronics

Area of Science:

• Materials Science
• Condensed Matter Physics
• Nanotechnology

Background:

• Van der Waals (vdW) dielectrics are crucial for enhancing nanoscale field-effect transistors (FETs) based on two-dimensional (2D) semiconductors.
• Ideal vdW dielectrics require high dielectric constants and suitable band alignment with 2D semiconductors, but high-quality options are limited.

Purpose of the Study:

• To screen van der Waals materials for potential use as dielectrics in 2D FETs.
• To identify promising dielectric candidates for MoS₂-based FETs and develop a machine learning model for accelerated screening.

Main Methods:

• Employed a topology-scale algorithm to screen 0D, 1D, and 2D vdW materials from the Materials Project database.
• Conducted high-throughput first-principles calculations to determine bandgaps and dielectric properties.
• Developed a two-step machine learning classifier with an active learning framework for dielectric screening.

Main Results:

• Calculated properties for 189 0D, 81 1D, and 252 2D vdW materials.
• Identified 9 promising dielectric candidates for MoS₂-based FETs.
• Discovered an additional 49 promising vdW dielectrics using the machine learning model.

Conclusions:

• The study provides a comprehensive list of potential vdW dielectrics for 2D FETs.
• A high-accuracy machine learning model was developed for efficient screening of vdW dielectrics.
• This work facilitates the future development and optimization of 2D FETs.