Fat band plot

[leostorm10]

Dear vasp users, I want to plot partial band exactly like partial dos. Could you please tell me how can I plot contributions of different atoms into band structure.Thanks in advance.

[juhL]

1st step: SCF cycle to get charge density
2nd step: non SCF calculation with ICHARGE = 11 and LORBIT = 11 on charge density of step 1
3rd step: visualization is possible in p4v

[leostorm10]

Thanks juhL, I did step 1 & 2 exactly as u wrote but I don't have p4v installed. I want to plot lm decomposed band using PROCAR file by writing a fortran program. But I donot know the lines in the PROCAR file. Even in the manual it is not described well. Could you please tell me how can I get the details of the PROCAR file.Thanks in advance.

[juhL]

.


<span class='smallblacktext'>[ Edited Wed Jun 15 2011, 06:50PM ]</span>

[juhL]

hi, hmm i've always used p4v for visualization, so i can't tell you without doubts how to write your own program, so it probably would be saver to trust other people, if they respond.

however, having a quick look at a PROCAR file, it seems quite obvious:

you have lists for each kpoint and each band; within every list, the values are separated into ions and s, p and d contributions

and that's all you need in my opinion

the only thing left to think about is the weight of the kpoints, and if (and how) to include it to your numbers, but since for "normal" k-point grids, all kpoints have the same weight, and all the weights add up to 1, this shouldn't be a problem and only a scaling factor


<span class='smallblacktext'>[ Edited Wed Jun 15 2011, 06:49PM ]</span>

[sktdebnath]

[quote="leostorm10"]Dear vasp users, I want to plot partial band exactly like partial dos. Could you please tell me how can I plot contributions of different atoms into band structure.Thanks in advance.[/quote]

its simple....take the contribution for the desired atom as well as desired orbital and divide it by total contribution (last column and last row). this way you will actually get a normalized contribution...write those contributions along with your band energy in a file. plot them !amazed !alien

[mforti]

Were you able to do it until now? I have been dealing with fat band plots for a few months, and I built an awk script to parse the procar file. for normal spin polarized calculations, and if lorbit=11, I understand that the general layout for the procar file is:
spin up{
k-point: i kx ky kz weight {
band 1: energy: E1 {
atom1: s py pz px dxy dyz dz2 dxz dx2 total
(all atoms)
total .....
}
{all bands}
}
{all kpoints}
}

{spin down}

and, to view it with a normal data plotter ( i. e. gnuplot, origin, etc) you need to parse these data into a table arranged like this (or watever you find useful) for every band:
(kpoint distance) Energy (s py pz px dxy dyz dz2 dxz dx2)_totals

This layout changes a little if you perform spin-orbit calculations. Instead of having spin up and spin down datablocks, for each band you have four blocks with m_tot, m_x, m_y and m_z contributions for all atoms.

Hope This helps! if this is wrong, I will also be grateful to anyone who can point me to the section in the manual where this is clearly explained!

About p4v, it is a great tool if you can make it run in your system, so I recommend you to try!