Segregation at interfaces in lightweight alloys towards tailored mechanical properties

Methods

Molecular statics and quasi-static simulations were performed using the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS). Various semi-empirical potentials including the Embedded-Atom Method (EAM) and Modified Embedded-Atom Method (MEAM) potentials were applied.

Results

Atomistic simulations were employed to assess dislocation and GB properties in Mg and its alloys, using various semi-empirical potentials. These assessments were then compared with experimental findings and first-principles calculations, highlighting the influence of chosen potentials on simulation outcomes. It was found that EAM potentials diverged significantly from experimental observations in terms of dislocation properties, particularly for pyramidal dislocations. In the context of the coincident site lattice (CSL) Σ7 grain boundary, the selection of interatomic potentials emerged as a critical factor determining the stability of minimum energy configurations. Here, T-type structures attained stability under MEAM potentials, while A-type structures were stable with EAM potentials. This study also included detailed examinations of ⟨0001⟩ symmetric tilt GBs, which demonstrated that, regardless of the potential used, type-1 GB energies were lower than those of type-2. When using selected potentials to analyze the generalized staking fault energy curves, the energy barrier for basal slip was overestimated in comparison to DFT data. In addition, an unexpected local minimum was evident for the prismatic slip system, which can be linked to the presence of a stable ⟨a⟩ screw dislocation on the prismatic plane, in contradiction to DFT results. Additionally, A strong correlation between the segregation energies and the local atomic environment was observed for all potentials, highlighting the significance of the minimization of elastic strain energy due to atomic size mismatch on solute segregation behavior.

Discussion

The observed discrepancies and correlations in our research underscore the pivotal role of choosing the right semi-empirical potentials for precise simulations. Our findings validate that MEAM potentials yield more accurate representations of dislocation behaviors than EAM potentials. For further investigations, we plan to employ MEAM potentials to delve into dislocation-grain boundary interactions. Additionally, we will conduct molecular dynamics simulations of grain boundary compression tests, aligning these simulations with corresponding experimental outcomes.

Statement:
In the previous period, we utilized only 100k Core-h from our 2.36 Mio Core-h allocation due to the phased nature of our three-year project, which integrates simulation with experimental work. The first year we first focused on benchmarking potentials for later studies. Next, simulations should be carried out based on the experimental outcome. Specifically, our research on Mg and its alloys’ grain boundaries required initial experimental efforts to grow crystals with precise grain boundaries. This process, compounded by the unexpected time needed to grow the Mg alloy with desired grain boundary. Moreover, an upgrade in our experimental apparatus – transitioning to a new furnace for better control of the crystal growth process – required significant time for calibration and parameter adjustment. These unforeseen challenges extended the duration of the growth experiments, thus delaying the simulation phase due to the lack of requisite experimental data for comparison. Consequently, the anticipated large-scale simulations were postponed, leading to the utilization of only 100k Core-hours during this period. For the new extension proposal, with the experimental hurdles now resolved, we are embarking on extensive large-scale simulations based on our experimental findings, which will need resources beyond 2 Mio Core-h.

Additional Project Information

DFG classification: 406 Materials Science
Software: LAMMPS, OVITO
Cluster: CLAIX