Abstract:Material characterization in nano-mechanical tests requires precise interatomic potentials for the computation of atomic energies and forces with near-quantum accuracy. For such purposes, we develop a robust neural-network interatomic potential (NNIP), and we provide a test for the example of molecular dynamics (MD) nanoindentation, and the case of body-centered cubic crystalline molybdenum (Mo). We employ a similarity measurement protocol, using standard local environment descriptors, to select ab initio configurations for the training dataset that capture the behavior of the indented sample. We find that it is critical to include generalized stacking fault (GSF) configurations, featuring a dumbbell interstitial on the surface, to capture dislocation cores, and also high-temperature configurations with frozen atom layers for the indenter tip contact. We develop a NNIP with distinct dislocation nucleation mechanisms, realistic generalized stacking fault energy (GSFE) curves, and an informative energy landscape for the atoms on the sample surface during nanoindentation. We compare our NNIP results with nanoindentation simulations, performed with three existing potentials -- an embedded atom method (EAM) potential, a gaussian approximation potential (GAP), and a tabulated GAP (tabGAP) potential -- that predict different dislocation nucleation mechanisms, and display the absence of essential information on the shear stress at the sample surface in the elastic region. We believe that these features render specialized NNIPs essential for simulations of nanoindentation and nano-mechanics with near-quantum accuracy.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is that in nano - mechanical testing, especially in nano - indentation testing, material characterization requires accurate inter - atomic potentials to calculate atomic energy and force in order to achieve an accuracy close to the quantum level. Traditional computational methods such as the finite element method (FEM), discrete dislocation dynamics (DDD) and molecular dynamics (MD) have limitations when simulating nano - mechanical tests, especially in capturing defect nucleation mechanisms and plastic deformation. Therefore, researchers have developed a neural - network - based inter - atomic potential (NNIP), aiming to improve the simulation accuracy of material behavior during nano - indentation by introducing specific configurations (such as generalized stacking - fault configurations and high - temperature configurations). These configurations can more accurately reflect the atomic environment of the material surface and dislocation core regions, so that NNIP can provide a more realistic and detailed energy landscape when simulating nano - indentation tests. Specifically, the goals of this paper include: 1. **Develop a high - precision NNIP**: Use machine - learning methods, especially neural networks, to construct an inter - atomic potential that can accurately describe the mechanical properties of materials at the nano - scale. 2. **Optimize the training data set**: Select and generate training data that can reflect the complex behavior of materials in nano - indentation tests, especially those data containing dislocation nucleation mechanisms and surface effects. 3. **Verify the effectiveness of NNIP**: By comparing with several existing inter - atomic potentials (such as EAM, GAP and tabGAP), evaluate the performance of the newly developed NNIP in simulating nano - indentation tests, especially the accuracy in predicting shear stress and dislocation nucleation mechanisms. Through the achievement of these goals, this research hopes to provide a new and more accurate method for the simulation of nano - mechanical tests, thereby promoting the design and development of new materials.

Neural Network Interatomic Potentials For Open Surface Nano-mechanics Applications

Molecular dynamics simulation on nanoindentation response of a copper workpiece

Modeling of Metal Nanoparticles: Development of Neural-Network Interatomic Potential Inspired by Features of the Modified Embedded-Atom Method

Towards Full Ab Initio Simulations of Mechanical Response of Ti-N Binary System under Extremely Operational Conditions Via Neural Network Potential

Development of neural network potential for MD simulation and its application to TiN

Atomistic simulations of nanoindentation in single crystalline tungsten: The role of interatomic potentials

A high-dimensional neural network potential for molecular dynamics simulations of condensed phase nickel and phase transitions

Pair vs many-body potentials: influence on elastic and plastic behavior in nanoindentation

High-Accuracy Neural Network Interatomic Potential for Silicon Nitride

Computer experiments on nano-indentation: A molecular dynamics approach to the elasto-plastic contact of metal copper

AI-dente: an open machine learning based tool to interpret nano-indentation data of soft tissues and materials

SG-NNP: Species-separated Gaussian Neural Network Potential with Linear Elemental Scaling and Optimized Dimensions for Multi-component Materials

Application of First-Principles-Based Artificial Neural Network Potentials to Multiscale-Shock Dynamics Simulations on Solid Materials

From Molecular Fragments to the Bulk: Development of a Neural Network Potential for MOF-5

LAMMPS Implementation of Rapid Artificial Neural Network Derived Interatomic Potentials

Inference and Prediction of Nanoindentation Response in FCC Crystals: Methods and Discrete Dislocation Simulation Examples

A deep learning interatomic potential developed for atomistic simulation of carbon materials

Fast and Sample-Efficient Interatomic Neural Network Potentials for Molecules and Materials Based on Gaussian Moments

Fast and accurate machine-learned interatomic potentials for large-scale simulations of Cu, Al and Ni

Atomistic simulations of the nanoindentation-induced incipient plasticity in Ni3Al crystal

Neural Network Potential with Multi-Resolution Approach Enables Accurate Prediction of Reaction Free Energies in Solution