My Community

Dear VASP Users,

What number should be set for the RWIGS flag for projected DOS calculation? My system is 1D polymer. According to the VASP manual, the simple way is to use the covalent radius of each species. But the OUTCAR show that the "Volume of Typ" is far below 100% for each species. What can I do? Thank you.

If you use PAW potentials, you don't need to specify RWIGS.
Otherwise, you may get a hint from the RWIGS value in the POTCAR file.

R.

[quote="Tian"]Dear VASP Users,

What number should be set for the RWIGS flag for projected DOS calculation? My system is 1D polymer. According to the VASP manual, the simple way is to use the covalent radius of each species. But the OUTCAR show that the "Volume of Typ" is far below 100% for each species. What can I do? Thank you.[/quote]
That is not your fault, just a consequence of living in a normal 3D space.
If you use covalent radii you will have no(or barely any) overlap such that you can never have 100%.
What you are trying to do is fill a cube with non-overlapping balls. And unless you are not working in 3 spatial dimensions you will not be able to fill your cube for the full 100% (so the 2 hints in the VASP manual are mutually exclusive, either you have nonoverlaping spheres(covalent radius) or you try to get 100% and you have overlapregions)

covalent radii for a bcc should give you 68% (If you only had 1 actual spatial dimension in your calculation then a bcc could give 100%)
for sc gives 52.36%
and for diamond you get 34% occupation.

So in the end:
If you use the covalent radius you will get an LDOS which is specific for the atom you are looking at. When you have overlap you might get peaks which are actually peaks from neighbouring different atoms(you'll see them appear when you set RWIGS to 150% of the covalent radius of your atoms )

Danny

please mind that if there is a substantial amount of vacuum in your cell, the sum of the 'atomic volumes will always be far off from 100%.
--) choose RWIGS such that neighboring atoms are touching, but non-overlapping.
--) LORBIT has to be set <10 in order to use the values specified for RWIGS in INCAR