UT theoretical chemistry code forum

Dear all,

I tried to use the Bader analysis via VASP to repeat the results of the given example (H2O gaussian cube）. I got the charge for H is 0.4254e, which does not agree with the value in the example (0.365e). Then I check the ACF.dat in the example and find the the X Y Z coordinate positions seem not to be in the right place of a real H2O (the distance between H and O is much larger than 1 Angstrom), while my OH bond distance is calculated to be 0.972 angstrom. Am I WRONG? I can't understand that.

Another question is that after I got the charge of H, i.e. 0.42 e, then can I use this value to calculate the dipole moment of H2O? It seems that definitely not. If not, then how can I utilize the bader analysis to get the "effective net charge ", saying similar to an experimental value of "ca. 0.33e"?
Can someone tell me how to do that?


Regards!

Using the H2O Gaussian cube file, I see a charge on H of 0.38, which is the same as in the ACF.dat file included in that example. For calculating a dipole moment, this means that the Bader partial charge should be +0.62 on each H and -1.24 on the O.

The cube files have a default distance unit of Bohr, which is why the bond lengths appear longer than you are expecting.

I'm curious about your statement regarding an experimental value for an atomic charge. While there are ways of quantifying partial charges experimentally, I do not believe that there is a unique way of defining or determining atomic charges.

Dear prof. Graeme

Thank you very much for the quick reply. More details are given about my calculation and former descriptions to make it clarified as follows:

For the first part, I still can't figure out whether my result is right or wrong. Now I have attached my INCAR , and the output files of ACF.dat from bader analysis after I followed the steps elaborated on your homepage (adding the core charges and valence charges, and then run the command "bader chargefile -ref charge_sum"). Grid numbers were increased and the charge values did't change too much (<=0.03e). Is my result right?

As for the experimental values of charge on H-atom, I got them from the website (http://www1.lsbu.ac.uk/water/water_molecule.html) and the Ref_1. On the website, it is said that 'The experimental values for gaseous water molecule are O-H length 0.95718 Å, H-O-H angle 104.474°' and in Ref_1 (TABLE II), it also uses the geometrical values and an Expt. value of moment 1.855 D. From this moment value, the charge on each H-atom can be derived reversely to be 0.33 e.

However, it is also explained on the website page that "The charge distribution depends significantly on the atomic geometry and the method for its calculation but is likely to be about -0.7e on the O-atom (with the equal but opposite positive charge equally divided between the H-atoms, i.e. +0.335) for the isolated molecule". And also I checked where the experimental values in Ref_1 come from, I found they are also calculated. Sorry for the carelessness and misunderstanding.

Yes, there are various ways to calculate charge distribution and the results can be quite different, as discussed in Ref_2. Then I have some questions, "how can we justify the results obtained from bade analysis?" To what extent can we say that our results are reasonable and credible, except for using the convergence test? Some questions may seem ill-considered since I am a newbie to learn this method. Please help me to figure these out.

Best regards.

Sorry, it seems that I am not allowed to attach files. All the above-mentioned files are listed as follows:
INCAR,

SYSTEM = H2O
PREC = High
ENCUT = 800
ISMEAR = 0
SIGMA = 0.01
EDIFF = 1E-6
EDIFFG= -0.01
NSW=300
IBRION=1
NFREE=2
ISIF=2
LWAVE=.FALSE.
LAECHG=.TRUE.

ACF.dat

# X Y Z CHARGE MIN DIST ATOMIC VOL
--------------------------------------------------------------------------------
1 7.5000000 7.5000000 7.5000000 7.1251935 0.6860490 43.1713113
2 8.0925000 6.7381950 7.5000000 0.4254407 0.1282684 7.2573570
3 8.0925000 8.2618050 7.5000000 0.4254407 0.1282684 7.2573570
--------------------------------------------------------------------------------
VACUUM CHARGE: 0.0239
VACUUM VOLUME: 3317.3140
NUMBER OF ELECTRONS: 8.0000

Ref_1, DOI: http://dx.doi.org/10.1103/PhysRevB.69.195404

Ref_2, DOI: 10.1002/jcc.20157