Phase field dislocation dynamics prediction of stacking fault widths in Cu-Al Alloys using generalised stacking fault energy from DFT and interatomic potentials
Abstract
The equilibrium width of intrinsic stacking faults is an important material parameter affecting key deformation phenomena. Here, we perform a comparative assessment of γ-surfaces generated using MD -- with a classical embedded-atom potential (ZW) and a machine-learned Gaussian Approximation Potential (GAP) -- and density functional theory (DFT), for Cu-Al alloys of varying Al content, and incorporate them as an energy term in PFDD simulations to obtain the equilibrium SFW. All three methods reproduce the lattice parameter and elastic moduli of Cu and Cu-Al alloys in reasonable agreement with each other and with experimental measurements from literature, but the MD-derived γ-surfaces, for both ZW and GAP, deviate substantially from DFT at the seven characteristic points, with the largest discrepancies at T and T1. When incorporated into PFDD, the DFT-derived γ-surfaces successfully predict an equilibrium SFW that increases with Al content and agrees well with standalone MD measurements and experimental values for pure Cu, whereas the MD-derived γ-surfaces do not yield an accurate SFW within PFDD, as the altered generalized stacking fault energy landscape causes the dissociating partials to become pinned soon after separation. These results show that DFT-derived γ-surfaces remain the more reliable route to alloy SFW prediction, and interatomic potentials must capture the shear path accurately to enable robust, efficient MD-PFDD modelling of stacking faults in alloys.
Related
- Revisiting quantum effects on dislocation glide in bcc metals from DFT calculations and machine-learning potentials
- Benchmarking machine-learning interatomic potentials using classical force fields
- Why Multi-Layer Message Passing Works: Completeness Theory for Graph Neural Network Interatomic Potentials
- MLIP Detective: Active Failure Mode Discovery Beyond Benchmark Scores for Machine-Learning Interatomic Potentials
- Cluster-based Structural Similarity for Dataset Visualization and Data Selection for Machine Learning Interatomic Potentials
BibTeX
@misc{varma2026phase,
title = {{Phase field dislocation dynamics prediction of stacking fault widths in Cu-Al Alloys using generalised stacking fault energy from DFT and interatomic potentials}},
author = {Arjun Varma R. and Ankosh D. Deshmukh and Sushil Kumar and Vaishali Shah and M. P. Gururajan},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.34420},
eprint = {2610.34420},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.34420},
}
RIS
TY - UNPB
TI - Phase field dislocation dynamics prediction of stacking fault widths in Cu-Al Alloys using generalised stacking fault energy from DFT and interatomic potentials
AU - Arjun Varma R.
AU - Ankosh D. Deshmukh
AU - Sushil Kumar
AU - Vaishali Shah
AU - M. P. Gururajan
PY - 2026
DA - 2026/10/25
PB - arXiv
DO - 10.48550/arxiv.2610.34420
UR - https://arxiv.org/abs/2610.34420
AB - The equilibrium width of intrinsic stacking faults is an important material parameter affecting key deformation phenomena. Here, we perform a comparative assessment of γ-surfaces generated using MD -- with a classical embedded-atom potential (ZW) and a machine-learned Gaussian Approximation Potential (GAP) -- and density functional theory (DFT), for Cu-Al alloys of varying Al content, and incorporate them as an energy term in PFDD simulations to obtain the equilibrium SFW. All three methods reproduce the lattice parameter and elastic moduli of Cu and Cu-Al alloys in reasonable agreement with each other and with experimental measurements from literature, but the MD-derived γ-surfaces, for both ZW and GAP, deviate substantially from DFT at the seven characteristic points, with the largest discrepancies at T and T_1. When incorporated into PFDD, the DFT-derived γ-surfaces successfully predict an equilibrium SFW that increases with Al content and agrees well with standalone MD measurements and experimental values for pure Cu, whereas the MD-derived γ-surfaces do not yield an accurate SFW within PFDD, as the altered generalized stacking fault energy landscape causes the dissociating partials to become pinned soon after separation. These results show that DFT-derived γ-surfaces remain the more reliable route to alloy SFW prediction, and interatomic potentials must capture the shear path accurately to enable robust, efficient MD-PFDD modelling of stacking faults in alloys.
ER -