LOCPROJ in empty sphere calculations

Questions will be moved to this forum when we consider them out of scope for support from our side: for instance when we do not have the necessary expertise to come up with an answer.
Another user still might, though!

Moderators: Moderator, Global Moderator

Post Reply
Message
Author
abhiyan_pandit
Newbie
Newbie
Posts: 6
Joined: Thu Jun 13, 2024 8:41 pm

LOCPROJ in empty sphere calculations

#1 Post by abhiyan_pandit » Thu Jun 13, 2024 9:49 pm

Dear VASP Experts/Users,

I am trying to do empty sphere calculations by using vasp_5.4.4 version with LOCPROJ tag (wiki/index.php/LOCPROJ). Following are my inputs.

$ more POSCAR
Na BCC structure
1.00000000000000
4.1928384359125666 -0.0000000000000000 -0.0000000000000000
-0.0000000000000000 4.1928384359125666 -0.0000000000000000
0.0000000000000000 0.0000000000000000 4.1928384359125666
Na
2
Direct
-0.0000000000000000 -0.0000000000000000 0.0000000000000000
0.5000000000000000 0.5000000000000000 0.5000000000000000

$ more INCAR
ISTART = 0
LCHARG = .T.
POTIM = 0.02
IALGO = 38
LREAL = Auto
SYSTEM = scf
ENCUT = 500
ISIF = 2
ICHARG = 2
LELF = .TRUE.
LAECHG = .TRUE.
IBRION = -1
PREC = Accurate
LOCPROJ = 1 2 : s p : Pr # <sites> : <functions-Ylm-specs> : <functions-radial-specs>
LOCPROJ = (0.25, 0.5, 0) : s : Pr
ISMEAR = 1
EDIFF = 1e-05
EDIFFG = -0.001
LWAVE = .T.
NSW = 0
SIGMA = 0.05
KSPACING = 0.2
PSTRESS = 0

Here I am trying to add an empty sphere at (0.25, 0.5, 0) and project s orbital on it.

However, I am getting an error message as below. Could you please see and suggest me the actual/correct way for doing it?
LPRJ_READER: ERROR: both POSCAR as well as POTCAR index of function: 9 not specified

Thank you so much in advance for your help.

Best,
Abhiyan Pandit

svijay
Global Moderator
Global Moderator
Posts: 73
Joined: Fri Aug 04, 2023 11:07 am

Re: LOCPROJ in empty sphere calculations

#2 Post by svijay » Fri Jun 14, 2024 3:28 pm

Dear Abhiyan Pandit,

In the LOCPROJ tag, you want to set a couple more inputs, say,

Code: Select all

LOCPROJ = (0.25, 0.5, 0): s : Pr 1 1
where the last two indices are specified here: wiki/index.php/LOCPROJ

The second to last index is that of the radial function of the nth projector (seen in the POTCAR file) and the last index is that of the atom type (1 in your example)

Sudarshan

abhiyan_pandit
Newbie
Newbie
Posts: 6
Joined: Thu Jun 13, 2024 8:41 pm

Re: LOCPROJ in empty sphere calculations

#3 Post by abhiyan_pandit » Fri Jun 14, 2024 6:59 pm

Dear Sudarshan,

Thank you so much for your reply.
So could you clarify more what is meant by 1 1 in 'LOCPROJ = (0.25, 0.5, 0): s : Pr 1 1'?
I simply want to add empty sphere(s), e.g. at (0.25, 0.5, 0), and project the orbitals onto the PAW projectors of the valence atomic sites (Na atoms) there. Please suggest me the correct way do it.
I was trying to following the instruction/example
'..........
To project the orbitals onto the PAW projectors of the valence atomic dxy-states of the atoms on sites 1 and 3, write

LOCPROJ = 1 3 : dxy : Pr
..........'
provided in the site: wiki/index.php/LOCPROJ.

Best,
Abhiyan

svijay
Global Moderator
Global Moderator
Posts: 73
Joined: Fri Aug 04, 2023 11:07 am

Re: LOCPROJ in empty sphere calculations

#4 Post by svijay » Mon Jun 17, 2024 2:22 pm

Dear Abhiyan Pandit,

The extra indices (1 1) from my previous answer are because you are projecting on a site that is not an atomic site. In this case, the code would need more information about the projectors. The way these are specified is:

Code: Select all

Pr [ nth-of-l [ species-number ] ]
Where the first index is that of the projector (information about which is supplied under the "Description" tag in the POTCAR, in y our example you have a choice of 1/2 for the s-state) and the second is of the species type, of which you only have one in your example.

Sudarshan

abhiyan_pandit
Newbie
Newbie
Posts: 6
Joined: Thu Jun 13, 2024 8:41 pm

Re: LOCPROJ in empty sphere calculations

#5 Post by abhiyan_pandit » Mon Jun 17, 2024 5:08 pm

Dear Sudarshan,

Now I got it, thank you so much for the helpful information.
And for more clarification with your info about 'Pr [ nth-of-l [ species-number ] ]' section in case I have multiple species system: would the species number be ranked in the order of them placed in POTCAR file? Is it possible to specify multiple species as well (if I want the gross contribution), and how?
Also, how can I supply multiple 'nth-of-l' projectors on a site if needed or should it be one at a time (e.g. for Mn with following 'Description' in the POTCAR file)?

Description
l E TYP RCUT TYP RCUT
1 -50.4666628 23 2.000
1 -1.3605826 23 2.000
2 -3.3744729 23 2.300
2 -4.7350555 23 2.300
0 -4.1331431 23 2.300
0 7.6656591 23 2.300
3 9.5240782 23 2.300


Best,
Abhiyan

svijay
Global Moderator
Global Moderator
Posts: 73
Joined: Fri Aug 04, 2023 11:07 am

Re: LOCPROJ in empty sphere calculations

#6 Post by svijay » Wed Jun 19, 2024 3:56 pm

would the species number be ranked in the order of them placed in POTCAR file?
Yes, it would be in the order of the species that you have in the POSCAR/POTCAR file
Also, how can I supply multiple 'nth-of-l' projectors on a site if needed or should it be one at a time (e.g. for Mn with following 'Description' in the POTCAR file)?
Yup, I would do them one at a time.

Sudarshan

abhiyan_pandit
Newbie
Newbie
Posts: 6
Joined: Thu Jun 13, 2024 8:41 pm

Re: LOCPROJ in empty sphere calculations

#7 Post by abhiyan_pandit » Wed Jun 19, 2024 7:02 pm

Hi Sudarshan,

I am still not sure: Following the 'Description' section of POTCAR (of Mn atom -> species 1) of my previous example, could you please show how can I supply the multiple projectors one at a time Vs all projectors in one line if I want (if possible)? I have the species number 1 for now.
Thank you so much for your help.

Best,
Abhiyan

abhiyan_pandit
Newbie
Newbie
Posts: 6
Joined: Thu Jun 13, 2024 8:41 pm

Re: LOCPROJ in empty sphere calculations

#8 Post by abhiyan_pandit » Thu Jun 20, 2024 4:39 am

Dear Sudarshan/VASP Developers/Experts,

In addition to my previous question, I also wanted to know about the data extraction/visualization from the LOCPROJ output file after the calculation (sample LOCPROJ and other relevant files are attached here for your better review). I am aware about the format of the file following wiki/index.php/LOCPROJ. But as each Kohn-Sham orbital is identified by the spin, k point, band index, eigenvalue, and occupation followed by the real and imaginary part of the projection; what is the common/more-meaningful way to plot/visualize these projections? I mean if we plot energy eigenvalue Vs real projection values of each orbitals (like we do for electronic DOS plots), how do we take into account all the parameters (spin, k point, band index, eigenvalue, and occupation)? OR Else? It would be great if you could share any other relevant references as well (if there are).
Thank you so much for your help.

Best,
Abhiyan
You do not have the required permissions to view the files attached to this post.

abhiyan_pandit
Newbie
Newbie
Posts: 6
Joined: Thu Jun 13, 2024 8:41 pm

Re: LOCPROJ in empty sphere calculations

#9 Post by abhiyan_pandit » Mon Jun 24, 2024 9:18 pm

Dear VASP Developers/Experts,

Could you please respond on my previous questions? We users rely on this VASP discussion forum as a main/resourceful platform (via instructions/suggestions and mutual knowledge sharing) in our research progresses that involves VASP implementation.
Again, thank you so much for your supports.

Best,
Abhiyan Pandit

svijay
Global Moderator
Global Moderator
Posts: 73
Joined: Fri Aug 04, 2023 11:07 am

Re: LOCPROJ in empty sphere calculations

#10 Post by svijay » Tue Jun 25, 2024 10:13 am

Dear Abhiyan Pandit,

You can supply as many options under the LOCPROJ tag, either through (for example)

Code: Select all

LOCPROJ = "1 : s : Hy 2
           1 3 : dxy : Ps 2 3"
or through separate LOCPROJ keys in the INCAR file (this option is only valid for VASP < 6).

For your second question: there are a few community driven tools for parsing and plotting LOCPROJ files and the exact nature of what you intend to plot depends on your specific application, especially because LOCPROJ contains a lot of information. I will move this post to the users-for-users category where you are likely to find a better response.

Sudarshan

Post Reply