Charge-density differencies between two cells?

[dyc_2008]

Dear all,

I want to compare the Charge-density differencies between two cells.
One is a pure.
The other one have a dopant atom in it. and the shape of the cell is bigger than the pure one because the dopant atom is more larger than the matrix atom.

I find some information at
http://cms.mpi.univie.ac.at/vasp-forum/ ... .php?4.586

1) make sure that all charge densities are written on the same grid
(same unit cell and FFT-mesh sizes)
2) subtract the different charge desities from one another (this
can be done by writing a small script or program)

However, I have two cells with different cell shape.
What can i do?

thank you !

[admin]

in this case, it is not possible to obrain charge density differences, because the grid points in CHGCAR are not compatible

[tlchan]

If the atomic positions and the unit cell are exactly the same, the only difference is the grid, then you can interpolate the grid for one of the calculations before subtracting the charge densities. Since the total charge density is typically quite smooth, the interpolation can be quite accurate.

If there are relaxations of atomic positions or unit cells, then it is more problematic. Usually, differences in charge densities are done to analyze charge transfer or changes in chemical bonding. If there are changes in either the atomic positions or the unit cell, then the charge density differences will mostly reflect the geometry changes, and the chemical effects will probably be obscured.

If your goal is to perform bonding analysis, maybe you can consider fixing the geometry after doping, or doing a self-consistent calculation without relaxation of the undoped system using the optimized doped geometry. This way, the two charge densities will be calculated on the same grid, and the charge density differences will be more meaningful.

[dyc_2008]

thank you admin and tlchan