I can confirm that 2 l-values per atom do not work with lapwso, but
should work with lapw1.
However, I do not have a fix for this and in fact I do not really plan
to introduce one, since I do not believe in putting U on two different l
values for the same atom is good physics.
For instance in your Gd example, the Gd-d states are too delocalized to
be treated with LDA+U. If you feel that GGA does not describe them well,
use mBJ+U.
I will, however, introduce a stop in lapwso so that the program does not
run (and neglects the first U value without telling you).
Peter Blaha
On 10/13/2016 10:47 PM, Laurence Marks wrote:
I am 99.9% certain that 2 U's for a given atom (orbital potentials) is
not supported in the version of lapwso that is available on the web.
Peter and/or Fabien may be able to provide you with a patched version
which will support 2 U's for a given atom.
On Thu, Oct 13, 2016 at 2:10 PM, Hung Yu Yang <yan...@bc.edu
<mailto:yan...@bc.edu>> wrote:
Dear wien2k users,
I am using wien2k 14.2 version to try to reproduce the GdSb
calculation in the following paper (see FIG.5 and FIG.6)<http:///>:
http://journals.aps.org.proxy.bc.edu/prb/abstract/10.1103/PhysRevB.74.085108
<https://urldefense.proofpoint.com/v2/url?u=http-3A__journals.aps.org.proxy.bc.edu_prb_abstract_10.1103_PhysRevB.74.085108&d=CwMFaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=Nln20MpCvs7TX6ZV_aaIcXr2drLPkFkUzlTTgYTq4M0&s=4dVfA821tmovnlSAKbvlMCWY_C6F_s3PuvHHPwNFuGM&e=>
I first did the calculation without SOC (LDA+spin-polarized+U), with
the following .indm file:
------------------ top of file: case.indm --------------------
-12. Emin cutoff energy
1 number of atoms for which density matrix is calculated
1 2 2 3 index of 1st atom, number of L’s, L1
0 0 r-index, (l,s)-index
------------------- bottom of file ------------------------
Similar changes were done in .inorb file. The result is satisfactory
in this case, as can be seen in the following link:
https://www.dropbox.com/s/fnqxvpgu3a8e3zg/GdSb_BS_woSOC_sp_d_f_dandf.pdf?dl=0
<https://urldefense.proofpoint.com/v2/url?u=https-3A__www.dropbox.com_s_fnqxvpgu3a8e3zg_GdSb-5FBS-5FwoSOC-5Fsp-5Fd-5Ff-5Fdandf.pdf-3Fdl-3D0&d=CwMFaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=Nln20MpCvs7TX6ZV_aaIcXr2drLPkFkUzlTTgYTq4M0&s=aEG4gTWkcOlecEaqj8m2d-U4M4Sikije_CcVaIc3cDo&e=>
In the two panels at the bottom, the gap around EF was open (from U
on d) and the f band was pushed down, which means the effects of U
on both d and f orbitals are well-considered.
Then I tried to do the calculation with SOC
(LDA+spin-polarized+U+SOC), and the result can be seen in the
following link:
https://www.dropbox.com/s/6cfbwu7yxcqxgsm/GdSb_SOC_bs.pdf?dl=0
<https://urldefense.proofpoint.com/v2/url?u=https-3A__www.dropbox.com_s_6cfbwu7yxcqxgsm_GdSb-5FSOC-5Fbs.pdf-3Fdl-3D0&d=CwMFaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=Nln20MpCvs7TX6ZV_aaIcXr2drLPkFkUzlTTgYTq4M0&s=k0tbwdjsUXEcsAwIOkbyBW92gdasKZqrZUWf0hfSe68&e=>
At the bottom right panel, although I tried to use the similar
setting to put U on both d and f, the effect of U only showed up on
d orbital (f orbital is not pushed down.) When I checked the
.outputorbup file, it shows
Calculation of orbital potential for spin block: up
Type of potential: LDA+U
Vorb applied to atom 1 orbit. numbers 2 3
Fully Localized Limit method
Atom 1 L= 2 U= 0.250 J= 0.000 Ry
Atom 1 L= 3 U= 0.600 J= 0.000 Ry
end of OP input
STRUCT file read
VSP read
Atom 1 L= 2 spin of potential; Lx, Ly, Lz= 0.000000 0.000000
-0.025894
Atom 1 L= 3 spin of potential; Lx, Ly, Lz= 0.000000 0.000000
0.003863
atom 1 L= 2 projection of L on M= -0.012830
atom 1 L= 3 projection of L on M= 0.158098
natom 1
No old potential found
Slater integrals F0, F2, F4 0.250 0.000 0.000 Ry
Ecorr 0.00011 Mult 1 Eldau 0.01465 Edc -0.03123 Tr(rho.V)
0.03016
:EORB: 0.00011466
Atom 1 spin up potential real part (Ry)
:VORBr 1_ 1 M= -2 0.10784 0.00000 0.00000 0.00000 -0.00757
:VORBr 1_ 1 M= -1 0.00000 0.11683 0.00000 0.00000 0.00000
:VORBr 1_ 1 M= 0 0.00000 0.00000 0.10180 0.00000 0.00000
:VORBr 1_ 1 M= 1 0.00000 0.00000 0.00000 0.11707 0.00000
:VORBr 1_ 1 M= 2 -0.00757 0.00000 0.00000 0.00000 0.11096
Potential imaginary part (Ry)
:VORBi 1_ 1 M= -2 0.00000 0.00000 0.00000 0.00000 0.00000
:VORBi 1_ 1 M= -1 0.00000 0.00000 0.00000 0.00000 0.00000
:VORBi 1_ 1 M= 0 0.00000 0.00000 0.00000 0.00000 0.00000
:VORBi 1_ 1 M= 1 0.00000 0.00000 0.00000 0.00000 0.00000
:VORBi 1_ 1 M= 2 0.00000 0.00000 0.00000 0.00000 0.00000
Slater integrals F0, F2, F4, F(6) 0.600 0.000 0.000 0.000 Ry
Ecorr 8.12337 Mult 1 Eldau 12.60579 Edc 12.95258 Tr(rho.V)
-1.99387
:EORB: 8.12337448
Atom 1 spin up potential real part (Ry)
:VORBr 1_ 1 M= -3 -0.29070 0.00000 0.00000 0.00000
0.00013 0.00000 0.00000
:VORBr 1_ 1 M= -2 0.00000 -0.28992 0.00000 0.00000
0.00000 0.00006 0.00000
:VORBr 1_ 1 M= -1 0.00000 0.00000 -0.28909 0.00000
0.00000 0.00000 0.00009
:VORBr 1_ 1 M= 0 0.00000 0.00000 0.00000 -0.28846
0.00000 0.00000 0.00000
:VORBr 1_ 1 M= 1 0.00013 0.00000 0.00000 0.00000
-0.28872 0.00000 0.00000
:VORBr 1_ 1 M= 2 0.00000 0.00006 0.00000 0.00000
0.00000 -0.28999 0.00000
:VORBr 1_ 1 M= 3 0.00000 0.00000 0.00009 0.00000
0.00000 0.00000 -0.29155
Potential imaginary part (Ry)
:VORBi 1_ 1 M= -3 0.00000 0.00000 0.00000 0.00000
0.00000 0.00000 0.00000
:VORBi 1_ 1 M= -2 0.00000 0.00000 0.00000 0.00000
0.00000 0.00000 0.00000
:VORBi 1_ 1 M= -1 0.00000 0.00000 0.00000 0.00000
0.00000 0.00000 0.00000
:VORBi 1_ 1 M= 0 0.00000 0.00000 0.00000 0.00000
0.00000 0.00000 0.00000
:VORBi 1_ 1 M= 1 0.00000 0.00000 0.00000 0.00000
0.00000 0.00000 0.00000
:VORBi 1_ 1 M= 2 0.00000 0.00000 0.00000 0.00000
0.00000 0.00000 0.00000
:VORBi 1_ 1 M= 3 0.00000 0.00000 0.00000 0.00000
0.00000 0.00000 0.00000
This seems to mean that the program actually read both U values on
the two orbital, but somehow it did nothing to the f orbital in the
band structure when SOC is included.
The way I did the calculation is that I first do a
LDA+SOC+spin-polarized scf calculation; then I made the .inorb and
.indm files, cp .indm .indmc, initso_lapw again, and run the
LDA+SOC+spin-polarized+U scf calculation. I tried to search the
mailing list but not much was done on the case in which one has U on
2 orbitals and includes SOC as well. I'm not sure this is a bug or I
am actually doing something wrong, could anyone help me with this
issue? Any information, comment or suggestion will be very helpful
and appreciated.
Yours sincerely,
Hung-Yu
--
Professor Laurence Marks
"Research is to see what everybody else has seen, and to think what
nobody else has thought", Albert Szent-Gyorgi
www.numis.northwestern.edu
<http://www.numis.northwestern.edu> ; Corrosion in 4D:
MURI4D.numis.northwestern.edu <http://MURI4D.numis.northwestern.edu>
Partner of the CFW 100% program for gender
equity, www.cfw.org/100-percent <http://www.cfw.org/100-percent>
Co-Editor, Acta Cryst A
_______________________________________________
Wien mailing list
Wien@zeus.theochem.tuwien.ac.at
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
SEARCH the MAILING-LIST at:
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
--
P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-165300 FAX: +43-1-58801-165982
Email: bl...@theochem.tuwien.ac.at WIEN2k: http://www.wien2k.at
WWW: http://www.imc.tuwien.ac.at/TC_Blaha
--------------------------------------------------------------------------
_______________________________________________
Wien mailing list
Wien@zeus.theochem.tuwien.ac.at
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
SEARCH the MAILING-LIST at:
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
