Planar averaged Hartree potential  [SOLVED]

Total energy, geometry optimization, DFT+U, spin....

Moderator: bguster

Locked
wxiaom86
Posts: 14
Joined: Tue Nov 29, 2016 6:28 pm

Planar averaged Hartree potential

Post by wxiaom86 » Tue Mar 07, 2017 3:00 am

Hello,

I am trying to calculate the ionization potential of the perovskites. So I construct a slab with thickness around 24 Bohr and vacuum of 30 Bohr. In order to align the VBM with the vacuum level, I printed the Hartree potential by using prtvha=1. Then I use macroave to get the planar averaged vhartree. To my surprise, vhartree across the surface direction is not consistent with the charge density oscillations, and vhartree at the vacuum region is not flat at all. Moreover, the vhartree obtained are huge numbers. Can anybody give me some help?

This the planar averaged Vhartree
vhartree.PNG
vhartree.PNG (79.69 KiB) Viewed 4786 times


The corresponding charge density
rho.PNG
rho.PNG (122.32 KiB) Viewed 4786 times


Following is my input file

enunit 1
ecut 30
occopt 1
nstep 300
nband 56
tolvrs 1.0d-10
iprcel 0
diemix 1.0
diemac 6
iscf 17
prtvha 1

kptopt 1
ngkpt 6 6 1
nshiftk 1
shiftk 0.0 0.0 0.0

# geo structure
acell 2*6.39 29.17 angstrom
angdeg 3*90
natom 12
ntypat 3
typat 1 2 3*3 1 2 3*3 2 3
znucl 82 55 53
xred 0.50000 0.50000 0.10953
0.00000 0.00000 0.00000
0.00000 0.50000 0.10953
0.50000 0.00000 0.10953
0.50000 0.50000 0.00000
0.50000 0.50000 0.32859
0.00000 0.00000 0.21906
0.00000 0.50000 0.32859
0.50000 0.00000 0.32859
0.50000 0.50000 0.21906
0.00000 0.00000 0.43812
0.50000 0.50000 0.43812


Best,
Xiaoming
UT

Julia
Posts: 7
Joined: Wed Aug 08, 2012 7:54 pm

Re: Planar averaged Hartree potential

Post by Julia » Mon Mar 13, 2017 10:57 am

Hello Xiaoming,

First of all, it seems that you used a large window for calculating the macroscopic average of the potential, is this what you wanted? A too large window would explain why the potential is not flat in the vacuum region, for most of the points in the vacuum region, the potential from the slab region still enters the average.

As for the oscillations in the charge density, it seems that you are comparing the planar average of the density (without running macroave) with the planar average of the potential. You do not see the oscillations in the potential, because the averaging got rid of them.

Julia
Last edited by Julia on Wed Mar 15, 2017 7:38 am, edited 1 time in total.

wxiaom86
Posts: 14
Joined: Tue Nov 29, 2016 6:28 pm

Re: Planar averaged Hartree potential

Post by wxiaom86 » Tue Mar 14, 2017 5:46 am

Julia wrote:Hello Xiaoming,

First of all, it seems that you used a large window for calculating the macroscopic average of the potential, is this what you wanted? A too large window would explain why the potential is not flat in the vacuum region, for most of the points in the vacuum region, the potential from the slab region still enters the average.

As for the oscillations in the charge density, it seems that you are comparing the planar average of the density (without running macroave) with the planar average of the potential. You do not see the oscillations in the potential, because the averaging got rid of them.

For the value of the Hartree potential, I think macroave expects the potential to be given in Hartree and converts it to eV. You asked abinit to output everything in eV (enunit 1) so the conversion was probably done twice (you need too check this yourself, as I am not 100% sure).

Julia


Hi Julia,

Thanks for you reply. For both of the potential and charge density, I am plotting the planar average not macroscopic average. So it is the PAV file from the output of macroave utility. I am using the same window for charge density and potential. However, I don't think the window will affect the PAV file.

Best,
Xiaoming

Julia
Posts: 7
Joined: Wed Aug 08, 2012 7:54 pm

Re: Planar averaged Hartree potential  [SOLVED]

Post by Julia » Tue Mar 14, 2017 10:19 pm

Hello Xiaoming,

I did not notice the names of the files you attached, you can forget my previous answer then.

My question now is, is the electron Hartree potential what you really wanted? If your goal is, for instance, to align the energy levels of the slab to vacuum, then you need the electrostatic potential (Hatree + local pseudo potential) or Vxc. These two potentials follow the oscillations of the charge density.

Julia

wxiaom86
Posts: 14
Joined: Tue Nov 29, 2016 6:28 pm

Re: Planar averaged Hartree potential

Post by wxiaom86 » Wed Mar 15, 2017 8:38 pm

Julia wrote:Hello Xiaoming,

I did not notice the names of the files you attached, you can forget my previous answer then.

My question now is, is the electron Hartree potential what you really wanted? If your goal is, for instance, to align the energy levels of the slab to vacuum, then you need the electrostatic potential (Hatree + local pseudo potential) or Vxc. These two potentials follow the oscillations of the charge density.

Julia


Hi Julia,

Yes, you are right. I am sorry. I meant the electrostatic potential instead of the hartree potential. I made a mistake. Thank you for your points. Now I print out the vpsp+vha, this time the potential oscillation looks consistent with the charge density.
electrostatic.PNG
electrostatic.PNG (133.43 KiB) Viewed 4747 times


Best,
Xiaoming

Locked