My Community

Dear All
I am trying to calculate the Born effective charge tensor for each atom of a system like LaMnO3. I didn't do this kind of calculations before. Can you please help me to do so?
I have checked the example of NaF given in the vasp manual, but didn't understand properly, so please help me to do so.
Regards
Rudra

is there anyone who can help?

also for more complex cells, please proceed as suggested in the manual:
1) first, do the 4 steps for the undistorted lattice (this is straighforward, I think): generate a converged CHGCAR file for the undistorted lattice (1) and calculate polarizations of the undistorted lattice as a point of reference (2-4).
2) Then, choose the polarization mode which you want to invesitgate by displacing the atoms which contribute this mode accordingly: In the NaF example given in the manual, the discussed mode corresponds to a displacement of the F atom of
0.0451Ã… along z. (it may be easier to choose carthesian input coordinates in POSCAR)
-- for this mode, generate a CHGCAR file to continue with (step 5).
-- proceed with steps 6-8 as you did for 2-4 for the undistorted structure.
please fix the origin of the dipole moment (DIPOL) such that its direction will not be reverted upon the displacement pattern you choose for the polarization mode you want to invesitgate.

Finally, please again proceed as discussed in the manual by calculating the averages of the electronic contributions of the distorted and undistorted structures, the electronic contribution of the dipole moment and finally the total change (by adding the ionic contribution). Depending on the chosen polarization mode, you will get the vector Delta P, (alpha components of Z) which have to be divided by the vector of the displacement vector (beta components of Z) you chose.
(please note that the NaF example given in the manual is a particularly easy one as the displacement of the F ion was done along z only, hence both the dipole moment vector as well as the displacement vector just have z components, yielding the Born effective charge tensor component Z 33

Thank you for your reply.
Following the process suggested by you I was trying to calculate the Polarization of PbTiO3.
First I have generated the CHGCAR file using the INCAR
LSCALU = .FALSE.
NSIM = 4
ISTART = 0
EDIFF = 0.00001
EDIFFG = 0.0001
IBRION = -1
NSW = 2
PREC = medium
ISMEAR = -5
SIGMA = 0.01
ISPIN = 1
VOSKOWN = 1
IALGO = 48
LDIAG = .TRUE.
NELM = 40
NELMIN = 4
NELMDL = -5
IMIX = 4
AMIX = 0.2
BMIX = .0001
ISYM = 2
LCHARG = .TRUE.
LELF = .FALSE.
LWAVE = .TRUE.
Then have done the next step as given in the manual just including
LBERRY = .TRUE.
IGPAR = 3
NPPSTR = 6
DIPOL = 0.5 0.5 0.25
in the INCAR.
I have done this whole process for both undistorted and distorted structure.
Output is:
(1)For undistorted structure
Expectation value term: <R>ev
<R> = ( 0.00000, 0.00000, 0.00000 ) electrons Angst
Berry-Phase term: <R>bp
<R> = ( 0.00000, 0.00000, 0.00000 ) electrons Angst
ionic term: <R>ion
<R> = ( 19.27001, 19.27001, -6.22800 ) electrons Angst
(2) For distorted structure
Expectation value term: <R>ev
<R> = ( 0.00000, 0.00000, 0.00355 ) electrons Angst
Berry-Phase term: <R>bp
<R> = ( 0.00000, 0.00000, -2.64500 ) electrons Angst
ionic term: <R>ion
<R> = ( 19.27001, 19.27001, -15.32173 ) electrons Angst

Then delta<R> = 11.739 electrons Angst
Volume of the cell = 61.671 Angst^3
Polarization = 304.97 microC/cm^2

Which is much higher than what it should me.
Please help me to figure out the mistake I have done.
Thanks in advance...
Rudra

you have to do the runs with IGPAR=1 and 2 for both structures as well to do the 3D avaeraging. Please follow the manual exactly!

Thank you for your reply.
I had repeated for both IGPAR=1 and 2, but as PbTiO3 has tetragonal structure and its polarization is along z axis, <R>bp is (0.0, 0.0, 0.0) for both IGPAR = 1 and 2. Thats why I didn't mention about it. I am really not understanding why am I getting so large polarization.
Please help me.
Thanks
Rudra