JARVIS-STM

JARVIS-STM provides DFT-calculated, Tersoff–Hamann–generated computational scanning tunneling microscope (STM) images for exfoliable two-dimensional (2D) materials to support structural analysis and machine-learning classification of lattice features.

Key Features:

• Extensive Database: Contains computational STM images for 716 distinct exfoliable 2D materials.
• DFT/Tersoff–Hamann Image Generation: Uses density functional theory calculations with the Tersoff–Hamann approach to produce simulated STM images.
• Bravais Lattice Analysis: Includes examples of five possible 2D Bravais lattice types along with their Fourier-transforms for structural characterization.
• Experimental Agreement: Computational STM images exhibit qualitative agreement with experimental STM images.
• Machine Learning Integration: Has been used to train convolutional neural networks to identify Bravais lattice structures from STM images, enabling high-throughput analysis.
• NIST-JARVIS Integration: Developed as part of the NIST-JARVIS framework.

Scientific Applications:

• Phase Identification: Identification of crystal phases in 2D materials via comparison of simulated and experimental STM images.
• Defect Analysis: Analysis of point and extended defects in 2D materials using computational STM contrasts.
• Lattice Distortion Examination: Examination of lattice distortions and symmetry via STM image patterns and Fourier analysis.
• Autonomous Experiment Workflows: Support for high-throughput and autonomous experimental workflows through ML-enabled classification of STM data.

Methodology:

Density functional theory (DFT) calculations using the Tersoff–Hamann method to generate STM images; computation of Fourier-transforms for lattice analysis; training of convolutional neural networks on STM images for Bravais lattice classification.