Machine-learning interatomic potential for W–Mo alloys

Abstract

In this work, we develop a machine-learning interatomic potential for WxMo1−x random alloys. The potential is trained using the Gaussian approximation potential framework and density functional theory data produced by the Vienna ab initio simulation package. The potential focuses on properties such as elastic properties, melting, and point defects for the whole range of WxMo1−x compositions. Moreover, we use all-electron density functional theory data to fit an adjusted Ziegler-Biersack-Littmarck potential for the short-range repulsive interaction. We use the potential to investigate the effect of alloying on the threshold displacement energies and find a significant dependence on the local chemical environment and element of the primary recoiling atom.

Modeling refractory high-entropy alloys with efficient machine-learned interatomic potentials: defects and segregation

“…Some further test results are provided in the Supplemental material, where we show that the tabGAP also reproduces the W-Mo alloy training dataset from Ref. [42] with similar accuracy as in Tab. I, verifying that the potential is accurate also for the binary alloys.…”

Simple machine-learned interatomic potentials for complex alloys

“…6, the test errors are separated by the different classes of structures in the test sets (where HEA refers to the MoNbTaVW high-entropy alloy). The figure also includes some pure-element test data from [41] as well as the liquids from the W-Mo dataset [12] and the W-Mo-Ta dataset [43], which were not included in the alloy test sets in the previous figures. The 2b+3b tabGAP in Fig.…”

“…Test data for pure elements [41] and liquid W-Mo and W-Mo-Ta from Refs. [12,43] are also included (which were not part of the alloy test sets in previous figures). The 2b+3b tabGAP is the one from Ref.…”

“…different phases or various crystallographic defects), molecules, or fixed-composition alloys and compounds [4][5][6][7]. This has left the challenges of developing ML potentials for materials of multiple elements much less explored, especially potentials that are transferable to arbitrary compositions of the included elements [8][9][10][11][12]. The main challenge for materials with many elements is the vast composition space that needs to be sampled by the training data.…”

Primary radiation damage in silicon from the viewpoint of a machine learning interatomic potential

“…Hence, the simulation capability that has been made available by ML potentials makes larger scale molecular dynamics simulations with DFT accuracy reachable [30][31][32][33][34][35][36][37][38][39]. That is why they have received a great amount of interest in the community and their implementation in different fields matures rapidly [40][41][42][43][44].…”