Unphysical effect in LOCPOT when using Dipole Correction?
Posted: Wed Sep 26, 2012 6:22 pm
Hi,
I am studying a Fe-slab with adsorbed atoms on the surface, among other properties I am studying the work function.
When I model this system using an asymmetric (adsorbed atom only on one side of the slab) slab and applying a dipole correction I find that when plotting the planar averaged potential as a function of z-value there is a "kink" in the potential in the center of the vacuum region (around 10 ? vacuum), i.e. the potential from one side of the slab does not converge to the same value as from the other. I find this to be clearly unphysical, and when modelling the same system with a symmetric slab (adsorbed atoms on both sides), or turning off the dipole correction this "kink" disappears. The converged value of the potential in the vacuum region is also slightly different. Has anyone seen the same behaviour? Comments? Glad for any help.
My solution so far is to model the system with a symmetric slab, with atoms on both sides, but to me it raises questions, especially when comparing results to results from calculations using the dipole correction.
I am using VASP 5.2.2, and can provide input files if called for.
Yours,
H?kan Hugosson
I am studying a Fe-slab with adsorbed atoms on the surface, among other properties I am studying the work function.
When I model this system using an asymmetric (adsorbed atom only on one side of the slab) slab and applying a dipole correction I find that when plotting the planar averaged potential as a function of z-value there is a "kink" in the potential in the center of the vacuum region (around 10 ? vacuum), i.e. the potential from one side of the slab does not converge to the same value as from the other. I find this to be clearly unphysical, and when modelling the same system with a symmetric slab (adsorbed atoms on both sides), or turning off the dipole correction this "kink" disappears. The converged value of the potential in the vacuum region is also slightly different. Has anyone seen the same behaviour? Comments? Glad for any help.
My solution so far is to model the system with a symmetric slab, with atoms on both sides, but to me it raises questions, especially when comparing results to results from calculations using the dipole correction.
I am using VASP 5.2.2, and can provide input files if called for.
Yours,
H?kan Hugosson