Abstract
By solving the problem of a periodic distribution of point defects in general anisotropic media, we give an alternative, more direct proof, of the relatively recent procedure that extracts dipole tensors from the stress acting on the cell of atomistic simulations performed under periodic boundary conditions. Moreover, we show that naive superposition of individual defect fields is not a solution of the problem, though correction terms can be identified; as a byproduct, analysis of the latter allows us to reveal a spurious contribution to the elastic interaction energy as calculated in current literature procedures, that therefore must be subtracted in order to obtain correct results.
Acknowledgements
The author is thankful to CNEA-GMAT project Aplicación de métodos de simulación clásicos y ab initio a aleaciones de interés nuclear, to the authors of Ref. [Citation11] for having made their source code available as supplementary material, and to the international community supporting free software development.
Notes
No potential conflict of interest was reported by the author.