My Community

I apologize for asking a question that seems to have been answered before, but I keep seeing the following as the answer:

AECCAR0 core charge density
AECCAR1 atomic AE (all-electron) charge density
(overlapping atomic charge density)
AECCAR2 AE charge denisty

However, AECCAR2 is not the AE charge density because the sum of charge in AECCAR2 is roughly equal to the sum of charge in CHGCAR (up to convergence) and the CHGCAR is the valence charge. If AECCAR2 was the AE charge density, it should include core charges, I'd think.

So:
1. Is AECCAR2 the valence charge corresponding to what would be an all electron calculation?
2. How does AECCAR2 differ from CHGCAR? Is it that CHGCAR includes compensation charges?
<span class='smallblacktext'>[ Edited ]</span>

1+2) AECCAR1 is a stratup file, so in general the values in there are nonsence. AECCAR2 is the same file after the calculation has run...so this file contains the AE valence density. It differs from the CHGCAR that the latter gives you the pseudo densities. For distances far from the nuclei this is exactly the same as AECCAR2, however close to the nuclei (<0.5A) they differ greatly, with the AECCAR showing the strong oscillations present in AE densities, while CHGCAR gives smooth density distributions, which integrate to the same number of electrons. (differences in the values obtained by integrating both files is due to the oscillations present near the nuclei in the AECCAR2, which require denser grids to get the same accuracy.)

Danny

Thanks for the reply. I was guessing that CHGCAR was the pseudo density + the compensation charge (see VASP "workshop" day 1). This could also explain why it integrates so well to the correct number of charges. But you're saying it's just the pseudo density, right?