Band decomposed charge densities

[cdt9f]

To generate charge densities around the fermi level we have used:

LPARD = TRUE
NBMOD = -3
EINT = -0.1 0.1
ISTART = 2
ICHARG = 1

As a sample calculation we used an Ag(111) surface with adsorbed O. However it appeared that a correct band decomposed charge density could not be obtained _unless_ the center layer of an odd-number-of-layers slab was positioned at z direct coordinate of 0.5 in the 'C' direction.

This seems ridiculous, but after extensive testing on this and another system, bare Pt(111), it seems to be the only way to get a reliable charge density. Can anyone comment on this?

Chris and Randall
University of Virginia

[admin]

the geometry of the slab should not affect the decomposition of the charge densities. Please check whether all other results of the runs compare (electronic structure, relaxed geometries,....).