Machine Vision Automated Chiral Molecule Detection and Classification in Molecular Imaging

Area of Science:

• Materials Science
• Nanotechnology
• Computational Chemistry

Background:

• Scanning probe microscopy (SPM) is crucial for atomic-scale imaging of surfaces and nanomaterials.
• Chiral molecular nanostructures are gaining importance in applications like catalysis and sensing.
• Current SPM image analysis is slow and prone to errors, necessitating automated solutions.

Purpose of the Study:

• To develop an efficient deep learning framework for analyzing SPM images.
• To automate the identification and chirality determination of chiral molecular nanostructures.
• To demonstrate the framework's accuracy and versatility in complex systems.

Main Methods:

• Utilized a state-of-the-art machine vision algorithm.
• Developed a one-image-one-system deep learning framework.
• Applied the framework to SPM images of supramolecular self-assemblies with distinct chiral patterns.

Main Results:

• The deep learning framework accurately detected molecular positions.
• Chirality of individual molecules within nanostructures was correctly labeled.
• Demonstrated high accuracy and versatility in analyzing complex chiral patterns.

Conclusions:

• The developed deep learning framework significantly enhances SPM image analysis.
• Machine learning shows great potential for automated recognition of complex SPM patterns.
• This approach facilitates the study of chiral molecular nanostructures for advanced applications.