UT theoretical chemistry code forum

Hello:
I have read a lot of threads in this list but still confusing about how to calculate the charge transfer using bader charges. Assuming the adsorbate is CO in my case, I have followed the step:

1. Use special flags to get the correct CHGCAR and calculate Bader charges of all atoms in Surface + Molecule
2. Get the charges of C and O: 2.8772 and 7.1309, so the charge transfer should be 2.8772+ 7.1309-10=0.0081 (10 is the formal charge of C(4) and O(6)) from surface to CO.

Is this method correct? However, the results in literature are that charge transfer from CO to surface. So what is probably the problem? Thank you very much!

That all seams reasonable. Presumably the transfer of charge to or from the surface should depend upon the electronegativity of the surface. Also, keep in mind that charge transfer is not a well-defined quantity apart from how it is defined. A Bader definition may not be the same as the experimental measure.

yes, I am aware of the problem of charge assignment based on plane wave calculations. However, it should affect the absolute charge quantity only. The loss or gain of electrons should not be influenced, is it a correct interpretation? If so, How can I judge the direction of electron flow?

My comment was not specific to plane wave calculations. My point is that a Bader calculation of charge transfer, even based upon an exact electronic structure calculation, may not correspond to an experimental measurement of charge transfer. In principle, the charge density could be measured experimentally and the same Bader analysis done -- this should agree with the calculation. But typically, charge transfer is measured experimentally with less direct methods, such as XANES, XES/XAS, or sometimes just with the assumption that it correlates with the CO vibrational frequency. So, what I am saying is that to compare with experiment, calculations should be done to evaluate the same quantity as the experiment. Simply saying 'charge transfer' is not a well-defined quantity apart from how it is measured.

So it implies that the analysis prior to experiment actually has no meaning at all? Without experimental results, the charge analysis could lead to diverse conclusions...

No, the Bader analysis can be very useful, especially when looking at trends. You just have to be careful when comparing to experiment or some other measure of charge transfer -- which is what you were asking about.

Er, last question! If there are no experimental and theoretical results available, how to validate my results that obtained from Bader analysis? I tried DOS in order to judge the charge transfer direction, but the results seem to contradict with the Bader charge. In my case, the CO adsorbate leads to a downshift of Fermi level, which should mean that CO is donor while the Bader charge analysis render it an acceptor…So what is the problem?

I think that the safest comparison is between Bader charge calculations, which will provide a relative measure of charge transfer in different situations. The issue of how charge transfer is defined is largely related to the absolute value.

You can also take the charge density of your entire system and subtract off the charge density of the isolated molecule and the surface. This will show you exactly where charge has moved in the interaction. You can even add up this charge difference in the Bader volumes.