Researchers from the Indian Institute of Technology Bombay, including Manish Sahoo, Akash Deshmukh, Yash Kokane, Jayaprakash H M, and Raghavan Ranganathan, have developed a new computational tool to accelerate the discovery of advanced materials for various industries, including energy. Their work, published in the journal Computational Materials Science, focuses on high-entropy alloys (HEAs), which are known for their exceptional properties and wide range of applications.
High-entropy alloys are materials that contain multiple principal elements in roughly equal proportions. These alloys have shown great promise in industries such as aerospace, energy, and chemical manufacturing due to their unique combination of thermodynamic, kinetic, and structural properties. However, exploring the vast compositional space of these alloys to discover new materials has been a challenge. Traditional experimental methods are time-consuming and costly, while computational methods have been limited by the lack of accurate interatomic potentials.
To address this challenge, the researchers developed a Moment Tensor Potential (MTP) using a machine learning-based approach. This potential was trained using the BFGS unconstrained optimization algorithm and a dataset that included various defect-induced configurations such as vacancies, dislocations, and stacking faults. The researchers employed an active learning scheme to dynamically retrain the potential, adding new training data when extrapolative configurations were encountered during non-equilibrium simulations.
The researchers thoroughly investigated the error metrics, equation of state, uniaxial tensile deformation, nano-indentation, and solid-liquid interface stability of the CoCrFeMnNi high-entropy alloy using their MTP potential. They found that their potential outperformed the popular Modified Embedded Atom Method (MEAM) potential in predicting physical properties. The MTP potential also demonstrated high computational speed and accuracy, making it a valuable tool for the rapid discovery of new alloy compositions and property measurements.
The researchers have made their potential publicly available by embedding it into the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) code. This development is expected to significantly accelerate the discovery and development of new high-entropy alloys for various industrial applications, including energy production and storage.
In the energy sector, high-entropy alloys could be used to develop more efficient and durable materials for energy conversion and storage devices, such as fuel cells, batteries, and supercapacitors. They could also be used to improve the performance and lifespan of energy generation and transmission infrastructure, such as turbines, pipelines, and power lines. The development of accurate and efficient computational tools like the MTP potential is a crucial step towards unlocking the full potential of these advanced materials.
This article is based on research available at arXiv.