[Wien] LDA+U & External Field

Post by Sherif Yehia
I would like to ask if the WIEN2k code allow for
LDA+U and an external field to be applied
simultaneously.

No WIEN cannot do this. However, it should be fairly simple to implement.

Basically you should run x orb twice and add the two corresponding
case.vorbup/dn files.

In addition make sure that when running lapw0 you have the input for a
magnetic field in case.inorb.

Regards

P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671 FAX: +43-1-58801-15698
Email: ***@theochem.tuwien.ac.at WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------

Pavel Novak

2005-07-06 19:13:18 UTC

Dear Sherif Yehia,

present WIEN2k implementation does not allow to run LDA+U and external
magnetic field simultaneously. You can circumvent the problem, however, if
you run orb two times - once for LDA+U, once for external field and then
use external field potentials in lapw0 and LDA+U potentials in lapwso. I
never tried that but it should work. An example of the UNIX script, you can
use to this end is attached. Note, however, that unlike in w2web I mix the
orbital potentials, not the density matrices (the script will be a bit
more simple if density matrices are mixed).

Regards
Pavel Novak

Post by Sherif Yehia
Hello All,
I am working on calculating Magnetic Anisotropy
Energy MAE in SmCo5 I hope to see effect of external
magnetic field.
I would like to ask if the WIEN2k code allow for
LDA+U and an external field to be applied
simultaneously.
In the case.inorb file has a switch which can either
call orbital polarization, LDA+U, or an external
magnetic field.(I hope there is another switch to
include both)
In case Wien2k does not do that I would welcome all
comments ,advice and precaution to develop better
understanding about the reason of not including both
effects.
Thanks all for efforts and time
Sherif Yehia
Physics Dept
Helwan Univ.
Cairo Egypt
__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product search
http://shopping.yahoo.com
_______________________________________________
Wien mailing list
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien

-------------- next part --------------
file=$1
num=$2
i=1
while [ $i -le $num ]
do
cp vorbup_ext $file.vorbup
cp vorbdn_ext $file.vorbdn
$HOME/wien2k/x -f $file lapw0
cp $file.clmsum $file.clmsum_old
cp $file.clmup $file.clmup_old
cp $file.clmdn $file.clmdn_old
cp vorbup_ext vorbup_ext_old
cp vorbdn_ext vorbdn_ext_old
cp vorbup_U vorbup_U_old
cp vorbdn_U vorbdn_U_old
cat $file.scf0>>$file.scf
$HOME/wien2k/x -f $file lapw1 -up
cat $file.scf1up>>$file.scf
$HOME/wien2k/x -f $file lapw1 -dn
cat $file.scf1dn>>$file.scf
cp vorbup_U $file.vorbup
cp vorbdn_U $file.vorbdn
$HOME/wien2k/x -f $file lapwso -orb -up
cat $file.scfso>>$file.scf
cp $file.energydum $file.energydumup
cp $file.energydum $file.energydumdn
$HOME/wien2k/x -f $file lapw2 -so -c -up
cat $file.scf2up>>$file.scf
$HOME/wien2k/x -f $file lapw2 -so -c -dn
cat $file.scf2dn>>$file.scf
$HOME/wien2k/x -f $file lapwdm -so -c -up
cat $file.scfdmup>>$file.scf
cp inorb_ext $file.inorb
cp vorbup_ext_old $file.vorbup_old
cp vorbdn_ext_old $file.vorbdn_old
$HOME/wien2k/x -f $file orb -up
cat $file.scforbup>>$file.scf
$HOME/wien2k/x -f $file orb -dn
cat $file.scforbdn>>$file.scf
cp $file.vorbup vorbup_ext
cp $file.vorbdn vorbdn_ext
cp inorb_U $file.inorb
cp vorbup_U_old $file.vorbup_old
cp vorbdn_U_old $file.vorbdn_old
$HOME/wien2k/x -f $file orb -up
cat $file.scforbup>>$file.scf
$HOME/wien2k/x -f $file orb -dn
cat $file.scforbdn>>$file.scf
cp $file.vorbup vorbup_U
cp $file.vorbdn vorbdn_U
$HOME/wien2k/x -f $file lcore -up
cat $file.scfcup>>$file.scf
$HOME/wien2k/x -f $file lcore -dn
cat $file.scfcdn>>$file.scf
mv *.dmatup dmatup
mv *.dmatdn dmatdn
$HOME/wien2k/x -f $file mixer
cat $file.scfm>>$file.scf
rm fo*
i=`expr $i + 1`
done

Sherif Yehia

2005-07-06 19:13:19 UTC

Dear Prof.Novak

Thank you very much for sending that wonderful and
genius script
I started as you advice before
1-SCF using GGA calculation
2-x symmetso
3-x kgen
4-run your Unix script jcontso
5-runsp_lapw -so
6-runsp_lapw -orb -so
all the above were done with the following smco5.inorb
1 1 0 nmod, natorb, ipr
BROYD 0.0 BROYD/PRATT, mixing
1 1 3 iatom nlorb, lorb
1 nsic 0..AFM, 1..SIC,
2..HFM
0.39 0.05 U J (Ry)

later for mixing Bext calculations
I ran vorb and used the following smco5.inorb
before using your Bext script

3 1 0 nmod, natorb, ipr
PRATT 1.0 BROYD/PRATT, mixing
1 1 3 iatom nlorb, lorb
4. Bext in T
0. 0. 1. Direction Bext in the
lattice vector

it gave no error but looking at the script
it needs these six files
1-vorbup_ext
2-vorbdn_ext
3-vorbup_U
4-vorbdn_ext
5-inorb_ext
6-inorb_U
which I didn't get after running vorb up/dn in my
working directory although it says the following in
case.outputorbup/dn
calculation of orbital potential for spin block: up
Type of potential: Interaction with Bext
Vorb applied to atom 1 orbit. numbers 3
end of OP input
Bext= 4.00000 T; muB*Bext= 0.17018E-04 Ry

Could you help me please and point my mistake with
the input for Bext.
I still with all the missing files tried to run
Bext
script. It develop four of the above missing files
(1-4) I noticed that all have the same data and no
change if I ran it for say n cycle

Till I hear you kindly advice and comments

Thanks very much for help ,support and your concern

my best wishes and regards

Sherif Yehia

Post by Pavel Novak
Dear Sherif Yehia,
present WIEN2k implementation does not allow to run
LDA+U and external
magnetic field simultaneously. You can circumvent
the problem, however, if
you run orb two times - once for LDA+U, once for
external field and then
use external field potentials in lapw0 and LDA+U
potentials in lapwso. I
never tried that but it should work. An example of
the UNIX script, you can
use to this end is attached. Note, however, that
unlike in w2web I mix the
orbital potentials, not the density matrices (the
script will be a bit
more simple if density matrices are mixed).
Regards
Pavel Novak

Post by Sherif Yehia
Hello All,
I am working on calculating Magnetic Anisotropy
Energy MAE in SmCo5 I hope to see effect of

external

Post by Sherif Yehia
magnetic field.
I would like to ask if the WIEN2k code allow for
LDA+U and an external field to be applied
simultaneously.
In the case.inorb file has a switch which can

either

Post by Sherif Yehia
call orbital polarization, LDA+U, or an external
magnetic field.(I hope there is another switch to
include both)
In case Wien2k does not do that I would welcome

all

Post by Sherif Yehia
comments ,advice and precaution to develop better
understanding about the reason of not including

both

Post by Sherif Yehia
effects.
Thanks all for efforts and time
Sherif Yehia
Physics Dept
Helwan Univ.
Cairo Egypt
__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product

Post by Sherif Yehia
http://shopping.yahoo.com
_______________________________________________
Wien mailing list

http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien

Post by Sherif Yehia
file=$1

num=$2
i=1
while [ $i -le $num ]
do
cp vorbup_ext $file.vorbup
cp vorbdn_ext $file.vorbdn
$HOME/wien2k/x -f $file lapw0
cp $file.clmsum $file.clmsum_old
cp $file.clmup $file.clmup_old
cp $file.clmdn $file.clmdn_old
cp vorbup_ext vorbup_ext_old
cp vorbdn_ext vorbdn_ext_old
cp vorbup_U vorbup_U_old
cp vorbdn_U vorbdn_U_old
cat $file.scf0>>$file.scf
$HOME/wien2k/x -f $file lapw1 -up
cat $file.scf1up>>$file.scf
$HOME/wien2k/x -f $file lapw1 -dn
cat $file.scf1dn>>$file.scf
cp vorbup_U $file.vorbup
cp vorbdn_U $file.vorbdn
$HOME/wien2k/x -f $file lapwso -orb -up
cat $file.scfso>>$file.scf
cp $file.energydum $file.energydumup
cp $file.energydum $file.energydumdn
$HOME/wien2k/x -f $file lapw2 -so -c -up
cat $file.scf2up>>$file.scf
$HOME/wien2k/x -f $file lapw2 -so -c -dn
cat $file.scf2dn>>$file.scf
$HOME/wien2k/x -f $file lapwdm -so -c -up
cat $file.scfdmup>>$file.scf
cp inorb_ext $file.inorb
cp vorbup_ext_old $file.vorbup_old
cp vorbdn_ext_old $file.vorbdn_old
$HOME/wien2k/x -f $file orb -up
cat $file.scforbup>>$file.scf
$HOME/wien2k/x -f $file orb -dn
cat $file.scforbdn>>$file.scf
cp $file.vorbup vorbup_ext
cp $file.vorbdn vorbdn_ext
cp inorb_U $file.inorb
cp vorbup_U_old $file.vorbup_old
cp vorbdn_U_old $file.vorbdn_old
$HOME/wien2k/x -f $file orb -up
cat $file.scforbup>>$file.scf
$HOME/wien2k/x -f $file orb -dn
cat $file.scforbdn>>$file.scf
cp $file.vorbup vorbup_U
cp $file.vorbdn vorbdn_U
$HOME/wien2k/x -f $file lcore -up
cat $file.scfcup>>$file.scf
$HOME/wien2k/x -f $file lcore -dn
cat $file.scfcdn>>$file.scf
mv *.dmatup dmatup
mv *.dmatdn dmatdn
$HOME/wien2k/x -f $file mixer
cat $file.scfm>>$file.scf
rm fo*
i=`expr $i + 1`
done

__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product search
http://shopping.yahoo.com

Pavel Novak

2005-07-06 19:13:19 UTC

Dear colleague,

in order to get the two input files for orb, you have to create them
separatelly and then, before running orb, copy corresponding file to
case.inorb:

cp inorb_ex smco5.inorb
x orb -up
cp smco5.vorbup vorbup_ex
cp inorb_U smco5.inorb
x orb -up
cp smco5.vorbup vorbup_U

and the same repeat for -dn.

I realized, however, that this is still not enough - before running
x lapwso -up -orb

it is necessary to sum vorbup_ex and vorbup_U (similarly dn) and copy the
sum to smco5.vorbup. To this end a simple program is to be written. I am
interested, so let me know more about your project and I could then do it.

Regards Pavel Novak

Post by Sherif Yehia
Dear Prof.Novak
Thank you very much for sending that wonderful and
genius script
I started as you advice before
1-SCF using GGA calculation
2-x symmetso
3-x kgen
4-run your Unix script jcontso
5-runsp_lapw -so
6-runsp_lapw -orb -so
all the above were done with the following smco5.inorb
1 1 0 nmod, natorb, ipr
BROYD 0.0 BROYD/PRATT, mixing
1 1 3 iatom nlorb, lorb
1 nsic 0..AFM, 1..SIC,
2..HFM
0.39 0.05 U J (Ry)
later for mixing Bext calculations
I ran vorb and used the following smco5.inorb
before using your Bext script
3 1 0 nmod, natorb, ipr
PRATT 1.0 BROYD/PRATT, mixing
1 1 3 iatom nlorb, lorb
4. Bext in T
0. 0. 1. Direction Bext in the
lattice vector
it gave no error but looking at the script
it needs these six files
1-vorbup_ext
2-vorbdn_ext
3-vorbup_U
4-vorbdn_ext
5-inorb_ext
6-inorb_U
which I didn't get after running vorb up/dn in my
working directory although it says the following in
case.outputorbup/dn
calculation of orbital potential for spin block: up
Type of potential: Interaction with Bext
Vorb applied to atom 1 orbit. numbers 3
end of OP input
Bext= 4.00000 T; muB*Bext= 0.17018E-04 Ry
Could you help me please and point my mistake with
the input for Bext.
I still with all the missing files tried to run
Bext
script. It develop four of the above missing files
(1-4) I noticed that all have the same data and no
change if I ran it for say n cycle
Till I hear you kindly advice and comments
Thanks very much for help ,support and your concern
my best wishes and regards
Sherif Yehia

Post by Sherif Yehia
Hello All,
I am working on calculating Magnetic Anisotropy
Energy MAE in SmCo5 I hope to see effect of

external

Post by Sherif Yehia
magnetic field.
I would like to ask if the WIEN2k code allow for
LDA+U and an external field to be applied
simultaneously.
In the case.inorb file has a switch which can

either

Post by Sherif Yehia
call orbital polarization, LDA+U, or an external
magnetic field.(I hope there is another switch to
include both)
In case Wien2k does not do that I would welcome

all

Post by Sherif Yehia
comments ,advice and precaution to develop better
understanding about the reason of not including

both

Post by Sherif Yehia
http://shopping.yahoo.com
_______________________________________________
Wien mailing list

http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien

Post by Sherif Yehia
file=$1

__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product search
http://shopping.yahoo.com
_______________________________________________
Wien mailing list
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien

Peter Blaha

2005-07-06 19:13:19 UTC

Post by Pavel Novak
I realized, however, that this is still not enough - before running
x lapwso -up -orb
it is necessary to sum vorbup_ex and vorbup_U (similarly dn) and copy the
sum to smco5.vorbup. To this end a simple program is to be written. I am

Alternatively it should be enough to include vorbup_ex in
x lapw1 -up -orb
but vorbup_U in
x lapwso -up -orb

vorb_ex should be small enough to neglect coupling of spins ?

P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671 FAX: +43-1-58801-15698
Email: ***@theochem.tuwien.ac.at WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------

Pavel Novak

2005-07-06 19:13:19 UTC

Post by Peter Blaha
Alternatively it should be enough to include vorbup_ex in
x lapw1 -up -orb
but vorbup_U in
x lapwso -up -orb

yes, that is clever, only if B_ext is in general direction, lapw1c should
be used.

Post by Peter Blaha
vorb_ex should be small enough to neglect coupling of spins ?

you are right, otherwise, if vorb_ex is large, nondiagonal vorbdnup must
also be considered

Pavel

Post by Peter Blaha
P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671 FAX: +43-1-58801-15698
--------------------------------------------------------------------------
_______________________________________________
Wien mailing list
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien

Sherif Yehia

2005-07-06 19:13:20 UTC

Dear Prof. Novak

Thank you very much for directing me to sum both
LDA+U and external field simultaneously your idea work
very nicely.

Now I have in my directory the four needed files to
be summed vorbup/dn_ex and vorbup/dn_U.I am very much
glad that you and prof. Blaha raised the problem of
coupling of spins that really very encouraging and
educative to me. (thanks to both of you).

Our (me and my colleague) interest in exploring the
dependence of the magnetic anisotropy on external
magnetic field have started at some point in our study
of ferromagnetic systems using a simple model that we
used to calculate magnetic properties given the
temperature dependence of both spontaneous
magnetization and anisotropy constants on temperature,
for example the papers :
JMMM 202,565(1999) ; ibid 212,195(2000) on Nd2Fe14B ;
and ibid 222, 368 (2000) on elemental Gd which I
coauthored with my colleague. A paper by F.Ono however
attracted our attention to the field-dependent
anisotropy in Fe ( F.Ono, JMMM. 54-57 (1986)
pp.907-908 ). The author discussed this dependence
theoretically in Refs 1,2 in the aforementioned paper.
According to his theory, part of that dependence comes
from a Fermi level dependent term which depends on
both of the average magnetic moment and anisotropy
constant at zero field and zero absolute temperature,
on the density of states of the minority spin subband
at EF and on other two terms which are generally of
weak temperature dependence. Therefore we ask the
questions about the possibility of incorporating the
influence of field in ab-initio calculations. Is our
question conceptual? Should one take into account the
possibility that a magnetic field of certain magnitude
and direction may change the symmetry of the crystal
under study?

SmCo5
needs LDA+U with the application of the
external field. I hope with the powerful and
accurate code we have and your stimulating idea I
can learn more about this permenat magnet.

Till I hear your advice and comments

wish you all the best

Sherif Yehia

Post by Pavel Novak
Dear colleague,
in order to get the two input files for orb, you
have to create them
separatelly and then, before running orb, copy
corresponding file to
cp inorb_ex smco5.inorb
x orb -up
cp smco5.vorbup vorbup_ex
cp inorb_U smco5.inorb
x orb -up
cp smco5.vorbup vorbup_U
and the same repeat for -dn.
I realized, however, that this is still not enough -
before running
x lapwso -up -orb
it is necessary to sum vorbup_ex and vorbup_U
(similarly dn) and copy the
sum to smco5.vorbup. To this end a simple program is
to be written. I am
interested, so let me know more about your project
and I could then do it.
Regards Pavel Novak

Post by Sherif Yehia
Dear Prof.Novak
Thank you very much for sending that wonderful

and

Post by Sherif Yehia
genius script
I started as you advice before
1-SCF using GGA calculation
2-x symmetso
3-x kgen
4-run your Unix script jcontso
5-runsp_lapw -so
6-runsp_lapw -orb -so
all the above were done with the following

smco5.inorb

Post by Sherif Yehia
1 1 0 nmod, natorb, ipr
BROYD 0.0 BROYD/PRATT, mixing
1 1 3 iatom nlorb, lorb
1 nsic 0..AFM,

1..SIC,

Post by Sherif Yehia
2..HFM
0.39 0.05 U J (Ry)
later for mixing Bext calculations
I ran vorb and used the following smco5.inorb
before using your Bext script
3 1 0 nmod, natorb, ipr
PRATT 1.0 BROYD/PRATT, mixing
1 1 3 iatom nlorb, lorb
4. Bext in T
0. 0. 1. Direction Bext in the
lattice vector
it gave no error but looking at the script
it needs these six files
1-vorbup_ext
2-vorbdn_ext
3-vorbup_U
4-vorbdn_ext
5-inorb_ext
6-inorb_U
which I didn't get after running vorb up/dn in my
working directory although it says the following

Post by Sherif Yehia
case.outputorbup/dn

Post by Sherif Yehia
Type of potential: Interaction with

Bext

Post by Sherif Yehia
Vorb applied to atom 1 orbit. numbers 3
end of OP input
Bext= 4.00000 T; muB*Bext= 0.17018E-04

Post by Sherif Yehia
Could you help me please and point my mistake

with

Post by Sherif Yehia
the input for Bext.
I still with all the missing files tried to run
Bext
script. It develop four of the above missing

files

Post by Sherif Yehia
(1-4) I noticed that all have the same data and no
change if I ran it for say n cycle
Till I hear you kindly advice and comments
Thanks very much for help ,support and your

concern

Post by Sherif Yehia
my best wishes and regards
Sherif Yehia

Post by Pavel Novak
Dear Sherif Yehia,
present WIEN2k implementation does not allow to

run

Post by Pavel Novak
LDA+U and external
magnetic field simultaneously. You can

circumvent

Post by Pavel Novak
the problem, however, if
you run orb two times - once for LDA+U, once for
external field and then
use external field potentials in lapw0 and LDA+U
potentials in lapwso. I
never tried that but it should work. An example

Post by Pavel Novak
the UNIX script, you can
use to this end is attached. Note, however, that
unlike in w2web I mix the
orbital potentials, not the density matrices

(the

Post by Pavel Novak
script will be a bit
more simple if density matrices are mixed).
Regards
Pavel Novak

Post by Sherif Yehia
Hello All,
I am working on calculating Magnetic

Anisotropy

Post by Sherif Yehia
Energy MAE in SmCo5 I hope to see effect of

external

Post by Sherif Yehia
magnetic field.
I would like to ask if the WIEN2k code allow

for

Post by Sherif Yehia
LDA+U and an external field to be applied
simultaneously.
In the case.inorb file has a switch which

can

Post by Pavel Novak
either

Post by Sherif Yehia
call orbital polarization, LDA+U, or an

external

Post by Sherif Yehia
magnetic field.(I hope there is another switch

Post by Sherif Yehia
include both)
In case Wien2k does not do that I would

welcome

Post by Pavel Novak
all

Post by Sherif Yehia
comments ,advice and precaution to develop

better

Post by Sherif Yehia
understanding about the reason of not

including

Post by Pavel Novak
both

product

Post by Pavel Novak
search

Post by Sherif Yehia
http://shopping.yahoo.com

_______________________________________________

Post by Sherif Yehia
Wien mailing list

http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien