[Wien] Spin-orbit coupling effect on band gap energy with LDA and GGA-PBE

[Peter Blaha]

I did not add RLOs at all.

For sure not for light elements like O and F.
This option is in particular for  semicore p states of heavy elements, 
but maybe for Bi-p it is ok.

In any case, if case.inso, in1, in2, in0 are identical for mbj-lda and 
mbj-pbe, the results must be identical.

Am 17.05.2024 um 22:24 schrieb Yakup Boran:
> Thank you for your valuable answer.
> Yes, Emax is the same for mbj_lda_so and mbj_pbe_so.
> 
> I checked your results and I am doing something wrong with the SO 
> calculation.
> 
> I've shared below the operation file for SO calculation. I would 
> appreciate it if you can tell me where I am doing wrong.
> 
> The Best
> 
> --------------------------------------------------------
> 
> hsn at hsn-Veriton-X4630G:~/WIEN2k/yakup/BiOF/so/mbj/nsp-gga-mbj-0-2/nsp$ 
> save_lapw -d uc
> 
> hsn at hsn-Veriton-X4630G:~/WIEN2k/yakup/BiOF/so/mbj/nsp-gga-mbj-0-2/nsp$ 
> init_so_lapw
> 
> The file nsp.in2c has been generated automatically
> 
> ---->Please select the direction of the moment ( h k l )
>                 (For R-lattice in R coordinates)(default 0 0 1):
> atom 1 is Bi
> atom 2 is O
> atom 3 is F
> 
> Select  atom-numbers (1,2,3) or "ranges of atoms" (1-3,9-12) (without 
> blanks)
> for which you would NOT like to add SO interaction
>   (default none, just press "enter" ):
> 
> For large spin orbit effects it might be necessary to include many more
> eigenstates from lapw1 by increasing EMAX in case.in1(c) - in case of
> MPI-parallel calculations with ELPA nband has to be increased instead.
> 
> ---->Please enter EMAX(default 5.0 Ryd):
> 
> The radial basis set for heavy atoms with p-semicore states is very
> limited. One can improve this by adding RLOs. Note: you MUST NOT add
> RLOs for atoms like oxygen,.... therefore the default is set to NONE
> ---->Add RLO for NONE, ALL, CHOOSE elements? (N/a/c) : c
> p-Energy parameters for Bi atom is :
>   1    0.30     0.0000 CONT 1
> 
> Would you like to add RLO? (Y/n)Y
> p-Energy parameters for O atom is :
>   1    0.30     0.0000 CONT 1
> 
> Would you like to add RLO? (Y/n)n
> p-Energy parameters for F atom is :
>   1    0.30     0.0000 CONT 1
> 
> Would you like to add RLO? (Y/n)n
>   Check the generated nsp.inso file (RLOs,...)
>   Check the generated nsp.in1 file (Emax and nband (if ELPA is used)
>   at the bottom of the file)
> 
> In spinpolarized case SO may reduce symmetry.
> 
> The program symmetso detects the proper symmetry and creates new struct and
> input files. (Note, equivalent atoms could become inequivalent in some 
> cases).
> 
> Do you have a spinpolarized case (and want to run symmetso) ? (y/N)N
> 
> Spinorbit is now ready to run.
> hsn at hsn-Veriton-X4630G:~/WIEN2k/yakup/BiOF/so/mbj/nsp-gga-mbj-0-2/nsp$ 
> run_lapw -i 200 -so -ec 0.00001 -cc 0.00001 -NI
> 
> On Fri, May 17, 2024 at 9:29 PM Peter Blaha <peter.blaha at tuwien.ac.at 
> <mailto:peter.blaha at tuwien.ac.at>> wrote:
> 
>     Using your struct file I tested the gaps. The results are exactly as
>     expected:
> 
>     psi11:/psi11/pblaha/test> grepline :gap '*scf' 1
>     in  6 files:
>     lda.scf::GAP (global)   :  0.243291 Ry =     3.310 eV (accurate
>     value if
>     proper k-mesh)
>     pbe.scf::GAP (global)   :  0.257682 Ry =     3.506 eV (accurate
>     value if
>     proper k-mesh)
>     mbj_lda.scf::GAP (global)   :  0.368924 Ry =     5.019 eV (accurate
>     value if proper k-mesh)
>     mbj_lda_so.scf::GAP (this spin):  0.336905 Ry =     4.584 eV (accurate
>     value if proper k-mesh)
>     mbj_pbe.scf::GAP (global)   :  0.368926 Ry =     5.020 eV (accurate
>     value if proper k-mesh)
>     mbj_pbe_so.scf::GAP (this spin):  0.336901 Ry =     4.584 eV (accurate
>     value if proper k-mesh)
> 
>     The lda gap is smallest, PBE a bit larger, and both mbj after lda or
>     mbj
>     after pbe give the same large gap.
>     Including SO yields of course also identical gaps, which are  by about
>     0.5 eV reduced.
> 
>     PS These calculations are just ewith   -prec 1
>     PPS: make a diff of your in1 files of mbj_lda_so and mbj_pbe_so. Is
>     Emax
>     the same ?
> 
> 
> 
> 
>     Am 17.05.2024 um 14:14 schrieb Yakup Boran:
>      > Dear Dr. Blaha,
>      >
>      > First of all, thank you for your time.
>      >
>      > DIS of all calculations are below.
>      >
>      > I checked the calculations again and everything seems correct to
>     me (of
>      > course, there might be something I am missing).
>      >
>      > I also share the struct file below for more detailed information. I
>      > would appreciate it if you could take a look at it when you have
>     a chance.
>      >
> 
>     -- 
>     --------------------------------------------------------------------------
>     Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
>     Phone: +43-1-58801-165300
>     Email: peter.blaha at tuwien.ac.at <mailto:peter.blaha at tuwien.ac.at>   
>     WIEN2k: http://www.wien2k.at <http://www.wien2k.at>
>     WWW: http://www.imc.tuwien.ac.at <http://www.imc.tuwien.ac.at>
>     -------------------------------------------------------------------------
> 

-- 
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-165300
Email: peter.blaha at tuwien.ac.at    WIEN2k: http://www.wien2k.at
WWW:   http://www.imc.tuwien.ac.at
-------------------------------------------------------------------------