smearing for projected dos on angular momentum channels

Total energy, DFT+U, BigDFT,...

Moderator: amadon

Post Reply
bmatt
Posts: 24
Joined: Tue Aug 27, 2013 9:46 pm

smearing for projected dos on angular momentum channels

Post by bmatt » Sat Nov 15, 2014 11:28 pm

Dear abinit users,
When using prtdos=3, with the proper variable declorations for natsph, iatsph, and ratsph a DOS output is created in my outputs directory (performing normally).

An issue occurs that I have found no variable to control. The output DOS contains the S P D F and G dos and integrated values. The problem exists with the data where sharp peaks appear when plotting, creating delta function-like peaks throughout the plot, lets say when plotting the D l=2 column with the energy. Is there any way to smear the pdos so that a smooth plot can be formed? (where the delta functions become broad)
With prtdos=1, you can use tsmear but prtdos=3 has no difference with different values of tsmear.

Is there a variable of interest, or is there a post processing tool that can help with this issue?
I have converged my kpt grid and that is not the issue.

I read somewhere that you can take a running average to help smear or broaden the data, but I have no idea how to do this. I would think that I am missing a variable because a pdso without an ability to smear or broaden seems a little rigid
Thanks for any suggestions or solutions

ilukacevic
Posts: 271
Joined: Sat Jan 16, 2010 12:05 pm
Location: Dept. of Physics, University J. J. Strossmayer, Osijek, Croatia
Contact:

Re: smearing for projected dos on angular momentum channels

Post by ilukacevic » Mon Nov 17, 2014 5:20 pm

Hi!

I think it occurs because prtdos=3 uses tetrahedron method as prtdos=2. Try using prtdos=1, which uses smearing method to calculate dos.

Best regards,

Igor

bmatt
Posts: 24
Joined: Tue Aug 27, 2013 9:46 pm

Re: smearing for projected dos on angular momentum channels

Post by bmatt » Mon Nov 17, 2014 10:52 pm

When prtdos = 1, you can use tsmear to control the smearing of the "total dos". I tried to declare the pdos variables when using prtdos = 1 but it just formed the total dos as an output.
Is there a way to obtain the dos for the S, P, D, F, and G orbitals where you can smear the results to obtain smooth plots? or do you have to use some post processing functions to achieve this?

ilukacevic
Posts: 271
Joined: Sat Jan 16, 2010 12:05 pm
Location: Dept. of Physics, University J. J. Strossmayer, Osijek, Croatia
Contact:

Re: smearing for projected dos on angular momentum channels

Post by ilukacevic » Tue Nov 18, 2014 10:53 am

Try to play with dosdeltae a little

http://www.abinit.org/documentation/helpfiles/for-v7.0/input_variables/vargs.html#dosdeltae

I think you'll have to trade the precision to get smoothness, but that's smth you'll have to decide yourself. I don't know of any post-processing tool for that.

Igor

Bibek Samal
Posts: 4
Joined: Fri Nov 08, 2019 9:52 am

Re: smearing for projected dos on angular momentum channels

Post by Bibek Samal » Thu Nov 28, 2019 6:07 pm

I also having same problems, along with that the numbers I am getting in y-axis for both in total dos and atomic contributions are substantially different. Why is that ?

Post Reply