UT theoretical chemistry code forum

Dear sir or madam

I employed vasp with bader to calculate the charge for a suface, using both PAW and LAECHG = .TRUE as you suggested. And made "bader CHGCAR -ref CHGCAR_sum"
However, the charge analysis results did not seem good.
You pointed out that selecting enough fft grid to converge the total charge density. I used a 210x54x180 grid for a 32x8x26 (angstrom) supercell and the charge density seems converged in bader. Is my grid fine enough ? Are there some emperical standar for the grid ?
Thank you.

lastleaf

Typically the default grid is somewhat course and a grid that is twice as fine is pretty good. But the best thing to do is to look for convergence of the Bader charges with respect to the grid density.

I increased the fft grid to 400x100x350. However, no satisfied results are obtained.
I want to estimate the charge transfer for several dimers on Ge surface. For two equivalent dimers on the surface, they are found to have different charge at the same atom, eg. one is 4.1,the other is 4.0. And their " MIN DIST" read from ACF.dat are different, one is 0.79 and the other is 0.66. It seems unreasonable.
What may casues these errors ?
Thanks.

There are two separate issues here. First, do the Bader charges converge with respect to the grid density. Second, do the results make sense.

We would be concerned if the analysis does not converge. If you can not get well defined charges with a fine grid, that is a problem.

If the results are not what you expect, this is a different issue. You are expecting the atoms in your dimer to be equivalent - are you sure that they are?