The physical properties of liquid alloy are crucial for many science fields. However, acquiring these properties remains challenging. By means of the deep neural network (DNN), here we presented a deep learning interatomic potential for the Ti–Ni–Cr–Al liquid system. Meanwhile, the thermophysical properties of the Ti–Ni–Cr–Al liquid alloy were experimentally measured by electrostatic levitation and electromagnetic levitation technologies. The DNN potential predicted this liquid system accurately in terms of both atomic structures and thermophysical properties, and the results were in agreement with the ab initio molecular dynamics calculation and the experimental values. A further study on local structure carried out by Voronoi polyhedron analysis showed that the cluster exhibited a tendency to transform into high-coordinated cluster with a decrease in the temperature, indicating the enhancement of local structure stability. This eventually contributed to the linear increase in the density and surface tension, and the exponential variation in the viscosity and the diffusion coefficient with the rise of undercooling.
11 September 2023 in Computational Materials Science
The thermophysical properties of liquid metals and alloys are crucial to explore the intrinsic mechanisms of the solidification process, glass formation, and fluid dynamics. The deep learning approaches have emerged as powerful tools in numerous scientific fields and exhibit extraordinary accuracy in the estimation of physical properties and structural characteristics for various materials. In this Tutorial, focusing on the thermophysical properties of liquid multicomponent alloys, deep learning methods, including both supervised learning and active learning, are introduced. Combined with the verification from electrostatic and electromagnetic levitation experiments, the influences of training parameters and methods on the accuracy to obtain interatomic potential by deep learning are revealed on the basis of deep neural network algorithm. As a result, this prediction method of liquid state properties for multicomponent alloys exhibited the dual advantages of high accuracy derived from density functional theory and low computational cost associated with empirical potential.
The establishment of reliable materials genome databases involving the thermophysical properties of liquid metals and alloys promotes the progress of materials research and development, whereas acquiring these properties imposes great challenges on experimental investigation. Here, we proposed a deep learning method and achieved a deep neural network (DNN) interatomic potential for the entire Ti–Ni–Cr–Al system from pure metals to quaternary alloys. This DNN potential exhibited sufficient temperature and compositional transformability which extended beyond the training and provided the prediction of the liquid structure and thermophysical properties for metallic materials with both density functional theory accuracy and classic molecular dynamics efficiency. The predicted results agreed well with the reported experimental data. This work opens a feasible way to address the challenges of rapidly and accurately acquiring thermophysical properties data for liquid pure metals and multicomponent alloys, covering a broad temperature range from superheated to undercooled state.