Events
Time: 1:30 p.m.
Place: IC 02/718
Vaclav Vitek, Department of Materials Science and Engineering, University of Pennsylvania, Philadelphia, USA
In bond-order potentials (BOPs) for transition metals only the bonding mediated by the d band is included explicitly and the covalent part of the cohesive energy is evaluated using Slater-Koster dd bond integrals. However, the effect of s electrons with orbitals centered on atoms neighboring the corresponding dd bond is not necessarily negligible. As shown by Nguyen-Manh et al. (Phys. Rev. Let. 85, 4136, 2000), this can be taken into account via screening of the dd bond integrals. In a recent paper of Lin at al. (Modell. Simul. Mater. Sci. Eng. 22, 034002, 2014) the dd bond integrals were determined using a projection scheme, developed by Madsen et al. (Phys. Rev. B 83, 4119, 2011), which utilizes the atomic orbitals that give the best representation of the electronic wave functions in DFT calculations. It was inferred that in this case the effect of s electrons was already included. In this talk we analyze this hypothesis by comparing studies employing BOPs with both unscreened and screened dd bond integrals. In all cases results are compared either with experiments or calculations based on the density functional theory (DFT). Studies of structures alternate to the BCC lattice, transformation paths that connect the BCC structure with FCC, simple cubic, BCT and HCP structures via continuously distorted configurations and calculations of γ-surfaces were all found to be insensitive to the screening of bond integrals. On the other hand, when the bond integrals are screened, formation energies of vacancies are improved, calculated phonon dispersion spectra reproduce the experimentally observed ones much better and dislocation core structure and dislocation glide are significantly different without and with screening of bond integrals. The latter lead to a much better agreement with available experiments. These findings suggest that the effect of s electrons on dd bonds, emulated by the screening of corresponding bond integrals, is the least significant when the lattice is distorted away from the ideal BCC structure homogeneously even if such distortion is large. Notwithstanding, when the distortion is local and inhomogeneous the impact of screening of the dd bond integrals is significant. In the studies presented such local inhomogeneities occur when phonons propagate through the lattice, at point defects and in the cores of dislocations.
