Page 1 of 1

Band structure calculations: PROCAR_OPT and Vasprun.xml disagree

Posted: Mon Jul 01, 2024 10:01 pm
by verena_neufeld1
Hi VASP team,

I've run band structure calculations first with an SCF calculation with a larger k mesh followed by a non SCF calculation to get the bands along the k point lines a) in the same calculation using KPOINTS as well as KPOINTS_OPT files and b) restarting non SCF from SCF using CHGCAR files (i.e. two calculations). I set LORBIT = 11. With option a), PROCAR_OPT and vasprun.xml have different projection info whereas with option b) PROCAR agrees with vasprun.xml (at least I think so from comparing a few numbers).

This can be reproduced by modifying the VASP Si band structure example, and setting LORBIT = 11, naming KPOINTS KPOINTS_OPT and using a KPOINTS file like

"""
K-Points
0
Monkhorst Pack
5 5 5
0 0 0
"""

Thanks!
Verena

Re: Band structure calculations: PROCAR_OPT and Vasprun.xml disagree

Posted: Tue Jul 02, 2024 7:37 am
by alexey.tal
Dear Verena,

Could you please provide all relevant input and output files for your calculation according to the forum guidelines.

Re: Band structure calculations: PROCAR_OPT and Vasprun.xml disagree

Posted: Wed Jul 03, 2024 10:18 pm
by verena_neufeld1
Thanks, I've attached the relevant files. As I mentioned, this is mostly copied from the vasp examples. All else that is needed is the Si PROCAR.

Re: Band structure calculations: PROCAR_OPT and Vasprun.xml disagree

Posted: Thu Jul 04, 2024 11:36 am
by alexey.tal
Thank you for reporting this. Indeed, it is a small bug in the routine that writes out the projection info into vasprun.xml. We will fix it in the next release.
For now you can apply the following patch to VASP6.4.3:

Code: Select all

diff --git a/src/linear_response.F b/src/linear_response.F
index dbd91519d..f17a908a8 100644
--- a/src/linear_response.F
+++ b/src/linear_response.F
@@ -1858,7 +1858,7 @@ CONTAINS
      ENDIF
 #endif
      IF (IO%IU6>=0) THEN
-       CALL XML_PROCAR(PAR, W_INTER%CELTOT, W_INTER%FERTOT, WDES_INTER%NB_TOT, WDES_INTER%NKPTS, LPAR,&
+       CALL XML_PROCAR(PAR_INTER, W_INTER%CELTOT, W_INTER%FERTOT, WDES_INTER%NB_TOT, WDES_INTER%NKPTS, LPAR,&
                      T_INFO%NIONP, WDES_INTER%NCDIJ, TAG="projected_kpoints_opt")
        OPEN(UNIT=99,FILE='PROCAR_OPT',STATUS='UNKNOWN')
        CALL WRITE_PROCAR(99,W_INTER,WDES_INTER,IO%LORBIT,LPAR,LMDIMP,PAR_INTER,PHAS_INTER,T_INFO,KPOINTS%EMIN,KPOINTS%EMAX)


Re: Band structure calculations: PROCAR_OPT and Vasprun.xml disagree

Posted: Mon Jul 08, 2024 9:09 pm
by verena_neufeld1
Thanks!