DFT+U

Message

apple · #1 Post by **apple** » Fri Oct 05, 2012 12:17 am

hi,
can someone comment if it is ok to assume two different Ueff values for the same type of atom (say Mn) in some Mn-based bulk oxide knowing that Mn may have different oxidation states (say +2 and +3) depending on how it is coordinated? Would such a calculation be wrong or too forceful?
Thanks.

kambiz · #2 Post by **kambiz** » Tue Oct 09, 2012 1:31 am

Hi,

If you have any properties, measured experimentally, you can compare your results to see how it works. Most papers, they usually consider one Ueff. Hopwever, if you define two same ion (Mn) in POTCAR you can define two Ueff values in INCAR.

Cheers

Kambiz

boris · #3 Post by **boris** » Tue Oct 09, 2012 1:15 pm

I think you would be right to do that. It has been shown in Fe compounds that Fe ions located on different lattice sites should have different values of U.

The most difficult part is to find the correct values for U depending on your Mn ions.

The fact that your Mn ions have different oxidation states will be reflected in the occupation matrices, which will be filled by different numbers of electrons.

You should first try with the same values of U, just to check if you get the correct trends, then try to improve the calculations by having different values.

Good luck!

apple · #4 Post by **apple** » Thu Oct 11, 2012 4:59 pm

Thanks for replying. I appreciate your comments.