Dear Lukasz,
I assumed you were dealing with Eu²⁺, where 7 4f electrons are
occupied. However, one of your 4f levels looks too close to the Fermi
energy (just 0.8 eV below) if you are using U = 0.6 Ry. Are you
working with Eu³⁺, where 6 4f electrons are occupied? Check your
spin moment to see if it has converged to the states that you
expected.
Best,
Yongbin
-------------------------
From: Wien <[email protected]> on behalf of Lee,
Yongbin [A LAB] via Wien <[email protected]>
Sent: Monday, July 22, 2024 9:01 AM
To: A Mailing list for WIEN2k users <[email protected]>
Cc: Lee, Yongbin [A LAB] <[email protected]>
Subject: Re: [Wien] 4f multiplet
Dear Lukasz
It appears that you didn't add spin-orbit coupling for the
calculation in your QTL. If that is correct (though it is a little
strange for me to calculate Eu without SO), f(3,3) and f(3,-3) should
be degenerate. Check f(3,-3) instead of f(3,3) to see if it is
working. Also, you can identify which bands have dominant
contributions for f(3,3)/f(3,-3) from your QTL file, at least for the
Gamma point (assuming your first k-point is Gamma). It looks like your
BAND 9 has dominant contributions from f(3,2) and f(3,-2). Check if
the band with f(3,3)/f(3,-3) is within your energy range.
Best,
Yongbin
-------------------------
From: Wien <[email protected]> on behalf of
pluto via Wien <[email protected]>
Sent: Monday, July 22, 2024 8:35 AM
To: A Mailing list for WIEN2k users <[email protected]>
Cc: pluto <[email protected]>
Subject: Re: [Wien] 4f multiplet
Dear Prof. Blaha, Prof. Marks,
Thank you for your remarks.
Regarding the fat-band plotting, I reproduced the Ylm(3,3) qtl problem
with EuO. Below I paste inq and qtlup files. The zipped outputqup file
is attached.
It seems all Ylm components are printed into case.qtlup file. Of
course
I can probably plot them using external code, but w2web is so
convenient
for quick checks.
In the case.inq I purposely calculated f and s states, in this order,
to
make sure that an extra column behind f-levels is printed out.
In case.insp, only using jtype = 9 makes a problem, nothing is
plotted.
Using jtype = 10 does not make a problem, s-character is plotted.
Best,
Lukasz
EuO-bulk-U-0.6Ry% more EuO-bulk-U-0.6Ry.insp
### Figure configuration
5.0 3.0 # paper offset of plot
10.0 15.0 # xsize,ysize [cm]
1.0 4 # major ticks, minor ticks
1.0 1 # character height, font switch
0.1 3 1 # line width, line switch, color
switch
### Data configuration
-5.0 5.0 2 # energy range, energy switch (1:Ry,
2:eV)
1 0.4852915091 # Fermi switch, Fermi-level (in Ry
units)
1 999 # number of bands for heavier
plotting 1,1
1 9 0.2 # jatom, jtype, size of heavier
plotting
EuO-bulk-U-0.6Ry% more EuO-bulk-U-0.6Ry.inq
-9.0 3.0 Emin Emax
1 number of atoms
1 1 0 0 iatom,qsplit,symmetrize,locrot
2 3 0 nL, l-values
EuO-bulk-U-0.6Ry% more EuO-bulk-U-0.6Ry.qtlup
EuO-bulk
LATTICE CONST.= 9.7208 9.7208 9.7208 FERMI ENERGY= 0.48846
141 < NMAT < 163 SPIN=2 NAT= 2 SO 0 KLmax 9
JATOM 1 MULT= 1 ISPLIT= 1
tot,f,(3;-3),(3;-2),(3;-1),(3;0),(3;1),(3;2),(3;3),s,
JATOM 2 MULT= 1 ISPLIT= 0 atom not selected for QTL calculation
BAND 1
-2.34799 1 0.97897 0.00000 0.00000 0.00000 0.00000 0.00000
0.00000
0.00000 0.00000 0.97896
-2.34799 2 0.00000
-2.34799 3 0.02103
-2.34801 1 0.97895 0.00000 0.00000 0.00000 0.00000 0.00000
0.00000
0.00000 0.00000 0.97895
-2.34801 2 0.00000
-2.34801 3 0.02105
-2.34805 1 0.97892 0.00000 0.00000 0.00000 0.00000 0.00000
0.00000
0.00000 0.00000 0.97892
-2.34805 2 0.00000
-2.34805 3 0.02108
-2.34811 1 0.97886 0.00000 0.00000 0.00000 0.00000 0.00000
0.00000
0.00000 0.00000 0.97886
-2.34811 2 0.00000
-2.34811 3 0.02114
-2.34820 1 0.97878 0.00000 0.00000 0.00000 0.00000 0.00000
0.00000
0.00000 0.00000 0.97878
...
BAND 9
0.42894 1 0.96574 0.96574 0.00000 0.45002 0.00000 0.06571
0.00000
0.45002 0.00000 0.00000
0.42894 2 0.00000
0.42894 3 0.03426
0.42867 1 0.96517 0.96517 0.03377 0.24701 0.18494 0.03374
0.18494
0.24701 0.03377 0.00000
0.42867 2 0.00000
0.42867 3 0.03483
0.42865 1 0.96413 0.96413 0.03281 0.27434 0.15808 0.03368
0.15808
0.27434 0.03281 0.00000
0.42865 2 0.00000
0.42865 3 0.03587
0.42857 1 0.96132 0.96132 0.02294 0.37307 0.06843 0.03244
0.06843
0.37307 0.02294 0.00001
0.42857 2 0.00000
0.42857 3 0.03868
...
On 2024-07-22 10:24, Peter Blaha wrote:
As far as I know it is impossible to get all those multiplet
stuctures
in DFT.
With respect to the partial DOS:
Does the case.qtlup/dn file look correctly and has all
decompositions ?
What says case.outputqup/dn ... ?
Am 19.07.2024 um 19:13 schrieb pluto via Wien:
Dear All,
I have been calculating bands for a ternary compound containing Dy.
Dy
4f exhibits a complex multiplet in photoemission. Is there a
correct
way to get something at least approximately correct from GGA+U? Or
is
it hopeless because of the correlated nature of the photoemission
final states (there are many more than 7 terms in the final state)?
I
was hoping to get something approximately correct to be able to
look
at the orbital decomposition.
I also noticed that w2web does not plot fat bands for Dy 4f(3,3)
orbital (Ylm decomposition calculated using qtl). An empty PS file
is
created. All other 4f Ylm orbital fat bands are plotted just fine.
Things do not depend on which other orbitals and in which order are
calculated by qtl, as can defined in case.inq. An example of
case.inq
file calculating s and f levels in Ylm basis:
-9.0 3.0 Emin Emax
1 number of atoms
1 1 0 0 iatom,qsplit,symmetrize,locrot
2 0 3 nL, l-values
Loc-rot matrix on Dy in case.struct is an identity:
Dy NPT= 781 R0=0.00001000 RMT= 2.5000 Z: 66.000
LOCAL ROT MATRIX: 1.0000000 0.0000000 0.0000000
0.0000000 1.0000000 0.0000000
0.0000000 0.0000000 1.0000000
Best,
Lukasz
_______________________________________________
Wien mailing list
[email protected]
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
SEARCH the MAILING-LIST at:
http://www.mail-archive.com/[email protected]/index.html
[1]
Links:
------
[1]
http://www.mail-archive.com/[email protected]/index.html
_______________________________________________
Wien mailing list
[email protected]
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
SEARCH the MAILING-LIST at:
http://www.mail-archive.com/[email protected]/index.html
