UT theoretical chemistry code forum

[/quote]You should add AECCAR0 and AECCAR2.[/quote]

There is a mistake in my post. I add AECCAR0 and AECCAR2.
But the result of Bader analysis on CHGCAR_sum is 1621.4448 core electrons

You won't be able to accurately calculate the total charge of the core electrons without an extremely fine grid. But I don't think this is what you want to do anyway. You probably want to do the Bader partitioning on the charge density sum (that you calculated) and the integration on the valance electrons:

bader CHGCAR -ref CHGCAR_sum

A little more information about this can be found here: http://theory.cm.utexas.edu/vtsttools/bader/vasp.php

"It is essential to increase the FFT grid (NGXF, NGYF, and NGZF) until the core charge converges to the expected value. "

From where I can see that the core charge converges to the expected value?

The standard output on the screen? or by

bader AECCAR0

??

This thread started many years ago and this statement about requiring convergence of the core charge is no longer necessary. Instead follow the instructions at:
http://theory.cm.utexas.edu/bader/vasp.php

Why is the total electrons not correct? I have set LAECHG=TRUE. and chosen PAW psudopotential. I also find that the total electrons chang with the cutoffenergy increaseing.

H C N Si
32 6 4 192

32+ 6*6+4*7+192*14=2784

but the total electron is 1674

why does this error happen?

Are you integrating the valence charge? I'm going to cut-and-paste from: http://theory.cm.utexas.edu/bader/vasp.php
Note that the partitioning is done on the total charge (CHGCAR_sum) but the integration is done on the normal CHGCAR file from vasp, containing only the valence electrons.

-------

By adding the LAECHG=.TRUE. to the INCAR file, the core charge is written to AECCAR0 and the valance charge to AECCAR2. These two charge density files can be summed using the chgsum.pl script:

chgsum.pl AECCAR0 AECCAR2
The total charge will be written to CHGCAR_sum.
The bader analysis can then be done on this total charge density file:

bader CHGCAR -ref CHGCAR_sum

-------

On the webpage: http://theory.cm.utexas.edu/bader/vasp.php, the last sentence says "One finally note is that you need a fine fft grid to accurately reproduce the correct total core charge. It is essential to do a few calculations, increasing NG(X,Y,Z)F until the total charge is correct. " How to increasing the NG(X,Y,Z)F ? Pls give some suggestions. Thanks!

set:
NGXF=xxx
NGYF=yyy
NGZF=zzz
in the INCAR file. Choose values which are (for example) twice as large as those reported in the OUTCAR from a standard calculation.

I have set NGX(Y,Z)F to be twice as the values in the standard calculations. After the calculation of "bader CHGCAR -ref CHGCAR_sum", I found the "NUMBER OF ELECTRONS" in ACF.dat is the number of valence electrons of the system. Is that correct? Where is the number of the total electron, including the core and valence electron. Thanks!

read this: http://theory.cm.utexas.edu/bader/vasp.php

You are integrating the valence charge density, so yes, the total number of electrons is the total number of valence electrons.

If you ran 'bader CHGCAR_sum' you would be analyzing the total charge, but the values would have large errors because of the cusps in the total charge density.

To graeme: you posted above (Mar 31, 2011) "..this statement about requiring convergence of the core charge is no longer necessary". Yet on Sep 29, 2011 you posted "Choose values [of NG(X,Y,Z)F] which are (for example) twice as large as those reported in the OUTCAR from a standard calculation." Are you saying that the tests changing NG(X,Y,Z)F are no longer necessary, or are they necessary?

The best procedure is to include the core charges in the total density and then integrate over the valence charge. This does not require sufficient resolution to resolve the cusps at the core; you should, however, increase the grid density to see convergence in the integrated valence charges.

This forum is really awesome!

[quote="graeme"]It is now possible to get core charges from the VASP code. This is important for Bader calculations in which most or all of the valance electrons participate with bonding.

Getting the core charge density from a PAW calculation requires updating to the latest version of the vasp code (4.6.31 08Feb07). With this code, the core and valance charge density will be written if the LAECHG = .TRUE. tag is set in the INCAR file. It is essential to increase the FFT grid (NGXF, NGYF, and NGZF) until the core charge converges to the expected value.

The core charge is written to the AECCAR0 and the valance to AECCAR2. These can be added using our chgsum.pl script. The bader analysis on this total charge density will not have any pseudopotential errors.[/quote]

HELLO~
What is the difference between:
1. "bader CHGCAR " directly, after self-consistence in VASP without LAECHG setting
2. set LAECHG=ture, then use AECCAR...blablabla
?

the first one also gives me the ACF.dat, which is sometimes the same with the second one.
What do the two ways differ? The easier way and the more complex way...

THANK YOU !!! ^_^

[quote="zwmiracle"][quote="graeme"]It is now possible to get core charges from the VASP code. This is important for Bader calculations in which most or all of the valance electrons participate with bonding.

Getting the core charge density from a PAW calculation requires updating to the latest version of the vasp code (4.6.31 08Feb07). With this code, the core and valance charge density will be written if the LAECHG = .TRUE. tag is set in the INCAR file. It is essential to increase the FFT grid (NGXF, NGYF, and NGZF) until the core charge converges to the expected value.

The core charge is written to the AECCAR0 and the valance to AECCAR2. These can be added using our chgsum.pl script. The bader analysis on this total charge density will not have any pseudopotential errors.[/quote]

HELLO~
What is the difference between:
1. "bader CHGCAR " directly, after self-consistence in VASP without LAECHG setting
2. set LAECHG=ture, then use AECCAR...blablabla
?

the first one also gives me the ACF.dat, which is sometimes the same with the second one.
What do the two ways differ? The easier way and the more complex way...

THANK YOU !!! ^_^[/quote]


OH .the former only contains valent electrons, and the latter contains core electrons too?
so if the test shows that my system is identital to these two ways, then i can use the easier one without worry?
sorry it is really not easy to get along with all your discussions....