Exploring the configuration spaces of surface materials using time-dependent diffraction patterns and unsupervised...

Area of Science:

• Computational materials science
• Surface science
• Nanotechnology

Background:

• Developing computational methods for exploring atomic configurations is crucial for advancing nanotechnology.
• Machine learning approaches require effective descriptors to encode structural features of candidate configurations.

Purpose of the Study:

• To propose and validate the use of time-dependent electron diffraction simulations for creating novel descriptors of surface material configurations.
• To demonstrate the capability of these descriptors in distinguishing subtle atomic arrangements and identifying metastable states.

Main Methods:

• Utilizing sub-femtosecond time-dependent electron diffraction simulations to capture sensitive atomic arrangement information.
• Developing descriptors based on the time-evolution of electron diffraction patterns.
• Applying these descriptors to search for metastable configurations in copper(111) and organic self-assembled monolayers.

Main Results:

• Demonstrated that time-dependent electron diffraction patterns are highly sensitive to atomic positions, even for configurations with identical symmetry.
• Successfully generated descriptors capable of distinguishing between different atomic arrangements within the unit cell.
• Identified metastable configurations in the studied surface material systems.

Conclusions:

• Time-dependent electron diffraction simulations offer a powerful new approach for generating atomic configuration descriptors.
• This method enables the discovery of complex and metastable surface material structures.
• The proposed technique holds significant potential for advancing computational materials science and nanotechnology.