Accelerating Density Functional Calculation of Adatom Adsorption on Graphene via Machine Learning.

Area of Science:

• Materials Science
• Computational Chemistry
• Nanotechnology

Background:

• Graphene exhibits unique properties, driving interest in its applications.
• Understanding graphene's interfacial properties is crucial for material design.
• Predicting these properties computationally can accelerate material discovery.

Purpose of the Study:

• To develop an automated workflow for predicting graphene interfacial properties.
• To guide the design of graphene-based materials, particularly composites.
• To accelerate the computational screening of graphene-reinforced metal matrix composites.

Main Methods:

• Utilized density functional theory (DFT) calculations for data generation.
• Implemented an automated routine for modeling adsorption structures and properties.
• Developed a machine learning model using engineered atomic features and a genetic algorithm for optimization.

Main Results:

• Achieved a mean percentage error below 35% for the machine learning model.
• Successfully predicted interfacial properties for graphene/magnesium composites.
• Demonstrated the workflow's efficacy in accelerating composite design.

Conclusions:

• The developed workflow efficiently predicts graphene interfacial properties.
• This approach offers a viable method for rapid design of graphene-reinforced composites.
• The general DFT acceleration routine has broad applicability in computational materials science.