Accurate simulations at the nanoscale depend on choosing appropriate interatomic potentials
(Nanowerk Spotlight) Over the past 20+ years, many science and engineering fields have seen a growing interest in the use of atomistic computer simulations such as molecular dynamics and Monte Carlo methods. This is especially true for nanoscience and nanotechnology, where more or less all advances require a detailed understanding of the manipulation of matter on an atomic, molecular, or supramolecular scale.
Nevertheless, there have been reservations over the accuracy of simulation results, notwithstanding the use of correct simulation procedure and conditions.
Improper choice of force field – known as interatomic potential – is one of the main concerns. The problem arises from the fact that a multitude of these interatomic potentials exist and it is very difficult to develop a universal interatomic potential that works appropriately for all applications.
"However, the robustness, accuracy, and validity of an atomistic simulations hinge on the appropriate choice of force fields," Seyed Moein Rassoulinejad-Mousavi, a doctoral student at the University of Missouri's College of Engineering, tells Nanowerk. "Force fields are key for modeling the interaction between atoms of a matter under study, and the challenge is to have an accurate force field working for any specific material at any desired temperature."
To serve this objective and make a benchmark as well as a shortcut for users to find their best force fields, Rassoulinejad-Mousavi together with Dr. Yijin Mao and Professor Yuwen Zhang from the Multiscale Thermal Transport Laboratory, have examined a number of force fields for materials that are popular in micro- and nanotechnologies.
For their studies, that are supported by a grant from National Science Foundation, the team used a repository of interatomic potentials from National Institute of Standards and Technology (NIST IPR) and the Sandia National Laboratories (LAMMPS) databases to study the accuracy of the force fields for frequently used materials at practical temperatures for real-world applications.
The Figure shows uniaxial tensile strain and mechanical strength of a copper cubic nanostructure with length of 20 nanometers in 240 picoseconds (240×10-12 seconds) (Image: University of Missouri) (click on image to enlarge)
"Based on our calculations, we concluded that inadequate choice of force field strongly affects the simulation results and gives rise to some inconveniences for calculations," says Rassoulinejad-Mousavi. "Some of the interatomic potentials seem to be useful and accurate for predicting one or two of the elastic constants or elastic modulus, not all of them."
"We also found that the elemental potentials that have been generated for a specific alloy or compound are not expected to necessarily work for all of the present species in the compound," he continues.
The novelty of this work is that the team showed for the first time, which among the many available models is accurate for any specific range of temperature or application.
The results presented are useful for interested researchers in the field of atomistic study of materials mechanical properties and increase the assurance of the users to see which interatomic potential fits well to their specific problem.
The scientists' ultimate goal is to investigate all single elements of the periodic table, binary (two elements) and trinary (three elements) and higher order (four or more elements) compounds, to eventually prepare a handbook of interatomic potentials for users and researchers in nanoscale simulations.