Machine Learning for the edge energies of high symmetry Au nanoparticles

TL;DR

This paper introduces a machine learning approach to accurately estimate edge energies of gold nanoparticles by decomposing total energy into contributions from various structural features, enabling efficient and reliable nanostructure analysis.

Contribution

The study develops a novel machine learning model that links total energy to geometrical features, providing a reliable, low-cost method to determine edge energies without complex edge length definitions.

Findings

Edge-energy density for (100)/(100) edges is 0.22 eV/Å.

Edge-energy density for (111)/(111) edges is 0.20 eV/Å.

Vertex energies are approximately 1 eV/atom.

Abstract

We present data-driven simulations for gold nanostructures, and develop a model that links total energy to geometrical features of the particle, with the ultimate goal of deriving reliable edge energies of gold. Assuming that the total energy can be decomposed into contributions from the bulk, surfaces, edges, and vertices, we use machine learning for reliable multi-variant fits of the associated coefficients. The proposed method of total energy calculations using machine learning produces almost ab-initio-like accuracy with minimal computational cost. Furthermore, a clear definition and metric for edge energy is introduced for edge-energy density calculations that avoid the troublesome definition of edge length in nanostructures. Our results for edge-energy density are 0.22 eV/{\AA} for (100)/(100) edges and 0.20 eV/{\AA} for (111)/(111) edges. Calculated vertex energies are about 1 eV/atom. The present method can be readily extended to other metals and edge orientations as well as arbitrary nanoparticle shapes.