Abstract
High-temperature thermal stability, elastic moduli and anisotropy are among the key properties, which are used in selecting materials for cutting and machining applications. The high computational demand of ab initio molecular dynamics (AIMD) simulations in calculating elastic constants of alloys promotes the development of alternative approaches. Machine learning concept grasped as hybride classical molecular dynamics and static first principles calculations have several orders less computational costs. Here we prove the applicability of the concept considering the recently developed moment tensor potentials (MTP), where moment tensors are used as material’s descriptors which can be trained to predict the elastic constants of the prototypical hard coating alloy, Ti0.5Al0.5N at 900 K. We demonstrate excellent agreement between classical molecular dynamics simulations with MTPs and AIMD simulations. Moreover, we show that using MTPs one overcomes the inaccuracy issues present in approximate AIMD simulations of elastic constants of alloys.
Highlights
Hard ceramic coatings are utilized to extend the wear resistance and lifetime of machining tools in the cutting zone subjected to excessive heat and mechanical load
It means that the knowledge of elastic constants at elevated temperature supports the design of high-strength, wear resistant industrial materials
As proof of concept, we show that moment tensor potentials (MTP) can be trained on ab initio molecular dynamics (AIMD) simulation results to obtain highly accurate elastic constants of a prototypical hard coating alloy, Ti0.5Al0.5N at 900 K
Summary
Hard ceramic coatings are utilized to extend the wear resistance and lifetime of machining tools in the cutting zone subjected to excessive heat and mechanical load. A bottleneck of all these calculations is the high computational de mand of the underlying AIMD simulations, especially in case of metallic alloys, where a large supercell model is needed in combination with dense k-point sampling in the Brillouin zone. This issue gets more severe in case of predicting elastic constants because of the need for accurate stresses. In accelerating computational schemes one does not want to add other errors on top of the density functional theory and the highly accurate (within 1 GPa) predictions of the temperature dependences of elastic. A comparison of calculated elastic properties using machine learned MTP with values calculated by highly accurate AIMD, as well as by approximate AIMD-based methods [17,31] shows that with the ma chine learned MTPs one achieves both, high efficiency and excellent accuracy of the simulated results
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
