Explicit k-points for older VASP version

Problems running VASP: crashes, internal errors, "wrong" results.


Moderators: Global Moderator, Moderator

Post Reply
Message
Author
mrm
Newbie
Newbie
Posts: 20
Joined: Mon Jul 07, 2008 2:24 am

Explicit k-points for older VASP version

#1 Post by mrm » Thu Aug 07, 2008 11:27 pm

Sorry, I also posted this at the "Physics Qs" forum. I did not realize that "Bugreports" is the right place for technical (non-physics) questions.

Dear All,

I am relatively new to VASP (not a complete beginner) and am trying to calculate the DOS and bandstructure of Si (for example). Following the VASP handson.pdf I was successful with the DOS but not with bandstructure. After much searching I found that the explicit form of the k-points file given in the handson.pdf is only good for VASP 4.6. However, I am presently using 4.4.5 and do not yet have access to 4.6.

Can someone please give me some guidance on how to enter the explicit k-points in the KPOINTS file so that I can generate the bandstructure? I know that following the sc calculation the IBZK file will contain the k-points.

1) Can I just copy IBZK into KPOINTS?
2) Should I include the tetrahedra?
3) Should I change the weighting for each k-point to one (1) or leave it as found in the IBZK file?
4) Will this include the high symmetry directions for the band structure?

Also, I am on a WinXP system (sorry) and do not have easy access to linux so p4vasp is not handy for viewing the bandstructure. Can someone please give hints on how to format the EIGENVAL data so I can plot it with gnuplot (or something similar)?

Any other suggestions you have would be much appreciated.

Thanks very much for any help you can provide.


<span class='smallblacktext'>[ Edited ]</span>
Last edited by mrm on Thu Aug 07, 2008 11:27 pm, edited 1 time in total.

admin
Administrator
Administrator
Posts: 2921
Joined: Tue Aug 03, 2004 8:18 am
License Nr.: 458

Explicit k-points for older VASP version

#2 Post by admin » Mon Aug 11, 2008 2:16 pm

1) the format of IBZKPT is compatible as kpoints input, but please mind that special k-points are needed to plot a bandstructure. Therefore please simply give the k-points along the high-symmetry directions of your BZ by hand, following the format of the IBZKPT file, if you don't want to upradeto vasp.4.6
2) this is not very reasonable, because the k-meshes you typically use to plot bandstructures don't yield a decomposition of the BZ into tetrahedra of comparable volume. Please use ISMEAR=0 or 1 instead
3) as a bandstructure run is just a post-processing run after having obtained a self-consistent chargedensity (ICHARG=11, no further electronic update), the weights of the k-points can be chosen to 1, they will not be needed to obtain a new chargedensity.
4) band structures are plotted along the high-symmetry lines of the BZ, whereas the k-points to generate the self-consistent chargedensities should be regularly spread over the IBZ (not just along the edges and high-symmetry lines). Therefore the k-meshes for scf/bandstructure runs will differ.
Last edited by admin on Mon Aug 11, 2008 2:16 pm, edited 1 time in total.

mrm
Newbie
Newbie
Posts: 20
Joined: Mon Jul 07, 2008 2:24 am

Explicit k-points for older VASP version

#3 Post by mrm » Wed Aug 13, 2008 8:31 pm

Dear Head-Admin,

Thank you for your help. I appreciate your time. I believe I understand now.
Last edited by mrm on Wed Aug 13, 2008 8:31 pm, edited 1 time in total.

Post Reply