Mechanical properties of Mo-Re alloy based on first-principles and machine learning potential function

Abstract

This study utilizes first principles calculations of density functional theory and Spectral Neighbor Analysis Potential (SNAP) machine learning potential to investigate the influence of rhenium concentration and temperature on the fundamental mechanical properties of Molybdenum-Rhenium alloy. The Mo1−xRex alloys (x = 0.0625–0.5) are constructed using a special quasi-random structure BCC model. The optimized geometries and lattices are used to calculate elastic constants and derivate mechanical parameters, including Bulk modulus, Young's modulus, Shear modulus, etc. The results show that with the increase of rhenium content, the mechanical properties of Mo1−xRex alloy are significantly improved, and higher than pure molybdenum, the best properties are reached at x(Re) = 0.3125. On the other hand, by analyzing the ratio of bulk modulus to shear modulus (B/G) and Poisson's ratio, the alloying of rhenium can also improve the ductility of molybdenum rhenium alloy. To address the challenge of calculating the high-temperature mechanical properties of Molybdenum-Rhenium alloy, a machine learning potential was developed within a training set called Spectral Neighbor Analysis Potential (SNAP). We accurately predicted the bulk modulus, shear modulus, Young's modulus, and Poisson's ratio of Molybdenum-Rhenium alloy over the temperature range of 300–1300 K. Additionally, we provided an accurate description of how temperature affects the mechanical properties and solubility of Molybdenum-Rhenium alloy. This approach aims to overcome the limitations associated with traditional methods and provide accurate predictions of the alloy's mechanical behavior at elevated temperatures.