vacancy formation energy

[atk_user]

Dear, all

I'm trying to calculate the oxygen vacancy formation in LaMnO3 bulk crystal by using above formula.

E (LaMnO3-x) + 1/2 E (O2) - E (LaMnO3)

E (LaMnO3-x) : total energy of crystal with single oxygen vacancy (with fully relaxation; cell and atom positions)
E (O2) : total energy of O2 molecule in 30 A cubic box
E (LaMnO3) : total energy of LaMnO3 perfect crystal (with fully relaxation; cell and atom positions)

However, the calculated formation energy is quite high compare to the result from VASP (plane-wave calculation) (same PBE functional). What's the problem? BSSE correction is needed due to basis sets?

How can I solve this problem? I need a detailed calculation procedure. Could you help me?

[Anders Blom]

BSSE is important, yes.

[atk_user]

How can I setup to consider BSSE for the vacancy formation energy. I found only an example of interlayer distance of graphite. Could you explain more details? I trying to calculate the BSSE correction but in manual there is an comment using GHOST atom instead of actually removing them from the configuration. (in the page 2)

http://quantumwise.com/documents/tutorials/latest/Grimme/Grimme.pdf

[Julian Schneider]

You can for example calculate the vacancy formation energy as described in Ref. 4 in the BSSE-tutorial (J. Theor. Comput. Chem. 11, 1261 (2012)) using a single oxygen atom as intermediate step.
Following the tutorial, when calculating the LaMnO3 system, the LaMnO3-x part of the system would then correspond to "layer1" in the tutorial, and the oxygen atom would correspond to "layer2".
Similarly, the two oxygen atoms in the O2-molecule atom would be "layer1" and "layer2", respectively.

[atk_user]

then,

<  the LaMnO3-x part of the system would then correspond to "layer1" in the tutorial, and the oxygen atom would correspond to "layer2".  >   <--  the energy from this calculation is correspond to the energy for LaMnO3-x ??

[atk_user]

I think the energy is correspond to the LaMnO3 !!

How about calculate the energy for the LaMnO3-x ??

After that could I using this formula?


E (LaMnO3-x) + 1/2 E (O2) - E (LaMnO3)

[Julian Schneider]

It's essentially all in the tutorial (cf. "Summarizing the calculations").
The total energy corresponds to LaMnO3, but the contributions (in energy, forces and stress) arising from the difference, e.g. between O being simulated inside the basis set of LaMnO3-x and as a single atom, are removed. Vice versa for LaMnO3-x .
Now, you have to optimize LaMnO3 with BSSE correction and get the total energy E(LaMnO3). Then you optimize and calculate  E (LaMnO3-x) and E(O) with the same settings but without BSSE.
Then you do the same thing for O2 (with BSSE) and O (without). and get the BSSE-corrected O2 binding energy from E(O2-binding) = E(O2)-2E(O).

Finally, you combine everything as shown in Ref. 4 in the tutorial to get the vacancy formation energy:
E(vac)= E(LaMnO3-x) - E(LaMnO3) + E(O) + 1/2 E(O2 - binding)