The ABINIT Users Mailing List ( CLOSED )

[Kris Andersen <andersen@physics.ucdavis.edu>]

Steven Homolya wrote:
> Q: What's the quickest way to get the differential electron density, 
> i.e., the difference between the Abinit calculated density and the
> sum of atomic densities?

It seems your question hasn't gotten much response. I've done this for 
clusters (i.e., finite systems: molecules, dots, etc.) using this procedure.

1. Calculate the total density.
2. For each atom in the cluster, calculate its density using the same 
parameters (ecut, acell, and rprim) to ensure the FFT grid sizes (ngfft) 
are the same.
3. Add up all the densities from step 2 and subtract the density from 
step 1.

Step 3 requires writing a simple program to read Abinit's (binary) *_DEN 
output files to add and subtract the density arrays. This can be done 
naively because the grid points match up. Abinit's multi-dataset mode 
simplifies step 2. I used this procedure to obtain Figures 3 and 4 in 
PRL 93, 246105 (2004).

Maybe somebody can suggest a procedure for solids?

--
Kristopher Andersen              http://yclept.ucdavis.edu/~andersen/
Department of Physics                    andersen@physics.ucdavis.edu
University of California at Davis                      (530) 754-9764