Re: [Wien] Energy convergence with spin-orbit

[Islam, Md F]

Thanks, prof Blaha. I was using TEMP broadening of 0.005 but I will try 
with a smaller number.

Fhokrul



From: wien-boun...@zeus.theochem.tuwien.ac.at 
[wien-boun...@zeus.theochem.tuwien.ac.at] On Behalf Of Peter Blaha 
[pbl...@theochem.tuwien.ac.at]
Sent: Saturday, February 13, 2016 3:26 AM
To: A Mailing list for WIEN2k users
Subject: Re: [Wien] Energy convergence with spin-orbit

Convergence, in particular at surfaces and with localized states at EF,
can be difficult.
In particular, the default TETRA method might give problems due to the
missing 3rd dimension and TEMP/TEMPS with some broadening (try as small
as possible (0.002) first, but eventually you need to increase it a little.
Also a too crude k-mesh can give problems and eventually you need to
increase it. In any case, for a magnetocrystalline anisotropy usually
huge k-meshes are necessary and careful testing is necessary, otherwise
you get meaningless numbers.

Am 12.02.2016 um 13:17 schrieb Islam, Md F:
> Dear Wien2k users,
>
>  I am trying to do a magnetic impurity calculation on a surface including 
> spin-orbit coupling.
> I set the energy and charge convergence to 0.1 and 0.001, respectively. 
> The calculation is
> well converged if magnetization is in-plane but for out of plane 
> magnetization it is not converging.
> While charge satisfies convergence criterion very well, the energy keeps 
> oscillating near the 3rd/4th
> decimal places.. I tried with different mixing (MSR1: 0.05 to 0.01) but 
> nothing seems to be working
> for about 2 months. For larger mixing, job crashes with QTL-B error. I would 
> appreciate if you can
> give me some suggestion to resolve the problem.
>
> I am also wondering if anyone has done any anisotropy calculation with 
> both self-consistent and
> non self-consistent methods and whether they give the same anisotropy energy.
>
>
> Thanks,
> Fhokrul
>
> ___
> 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
>

--
--
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.atWIEN2k: http://www.wien2k.at
WWW:   http://www.imc.tuwien.ac.at/staff/tc_group_e.php
--
___
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
___
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

Re: [Wien] Energy convergence with spin-orbit

[Peter Blaha]

Convergence, in particular at surfaces and with localized states at EF, 
can be difficult.
In particular, the default TETRA method might give problems due to the 
missing 3rd dimension and TEMP/TEMPS with some broadening (try as small 
as possible (0.002) first, but eventually you need to increase it a little.
Also a too crude k-mesh can give problems and eventually you need to 
increase it. In any case, for a magnetocrystalline anisotropy usually 
huge k-meshes are necessary and careful testing is necessary, otherwise 
you get meaningless numbers.


Am 12.02.2016 um 13:17 schrieb Islam, Md F:

Dear Wien2k users,

 I am trying to do a magnetic impurity calculation on a surface including 
spin-orbit coupling.
I set the energy and charge convergence to 0.1 and 0.001, respectively. The 
calculation is
well converged if magnetization is in-plane but for out of plane magnetization 
it is not converging.
While charge satisfies convergence criterion very well, the energy keeps 
oscillating near the 3rd/4th
decimal places.. I tried with different mixing (MSR1: 0.05 to 0.01) but nothing 
seems to be working
for about 2 months. For larger mixing, job crashes with QTL-B error. I would 
appreciate if you can
give me some suggestion to resolve the problem.

I am also wondering if anyone has done any anisotropy calculation with both 
self-consistent and
non self-consistent methods and whether they give the same anisotropy energy.


Thanks,
Fhokrul

___
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



--
--
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.atWIEN2k: http://www.wien2k.at
WWW:   http://www.imc.tuwien.ac.at/staff/tc_group_e.php
--
___
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

[Wien] Energy convergence with spin-orbit

[Islam, Md F]

Dear Wien2k users,

I am trying to do a magnetic impurity calculation on a surface including 
spin-orbit coupling. 
I set the energy and charge convergence to 0.1 and 0.001, respectively. The 
calculation is 
well converged if magnetization is in-plane but for out of plane magnetization 
it is not converging. 
While charge satisfies convergence criterion very well, the energy keeps 
oscillating near the 3rd/4th 
decimal places.. I tried with different mixing (MSR1: 0.05 to 0.01) but nothing 
seems to be working 
for about 2 months. For larger mixing, job crashes with QTL-B error. I would 
appreciate if you can 
give me some suggestion to resolve the problem.

   I am also wondering if anyone has done any anisotropy calculation with both 
self-consistent and 
non self-consistent methods and whether they give the same anisotropy energy.


Thanks,
Fhokrul

___
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

Re: [Wien] Energy convergence with spin-orbit

[Laurence Marks]

I have only done a few -so calculations, but from the list they seem to
lead to problems. A few suggestions:

a) Check for user error when you setup the in plane calculation. Have you
used the right symmetry reduction? This seems to be the most common issue.

b) Check that you are using the patches for issues in initso that have been
posted.

c) Look if your calculation is oscillating or just veery slowly
converging. You can do a grep on some of the moments such as :ORB, :POM,
:SPI in case.scf. I have seen these only slowly converge as the occupied
states change.

d) Think carefully about the physics of your problem. Some time ago I did a
test LDA+U calculation for a Ni (metal) surface. It did not converge, which
was OK. In Wien2k you are solving a fixed-point problem, and always
assuming that such a solution exists. I do not believe that anything
guaruntees that a solution exists in all cases. ( This is different from
doing a variational calculation.) Maybe with your distances your impurities
want to have an AFM ordering in-plane, but you are preventing this.

These are just guesses, without more information it is hard to know.

---
Professor Laurence Marks
Department of Materials Science and Engineering
Northwestern University
http://www.numis.northwestern.edu
Corrosion in 4D http://MURI4D.numis.northwestern.edu
Co-Editor, Acta Cryst A
"Research is to see what everybody else has seen, and to think what nobody
else has thought"
Albert Szent-Gyorgi
On Feb 12, 2016 06:19, "Islam, Md F"  wrote:

> Dear Wien2k users,
>
> I am trying to do a magnetic impurity calculation on a surface
> including spin-orbit coupling.
> I set the energy and charge convergence to 0.1 and 0.001,
> respectively. The calculation is
> well converged if magnetization is in-plane but for out of plane
> magnetization it is not converging.
> While charge satisfies convergence criterion very well, the energy keeps
> oscillating near the 3rd/4th
> decimal places.. I tried with different mixing (MSR1: 0.05 to 0.01) but
> nothing seems to be working
> for about 2 months. For larger mixing, job crashes with QTL-B error. I
> would appreciate if you can
> give me some suggestion to resolve the problem.
>
>I am also wondering if anyone has done any anisotropy calculation with
> both self-consistent and
> non self-consistent methods and whether they give the same anisotropy
> energy.
>
>
> Thanks,
> Fhokrul
>
> ___
> 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
>
___
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

Re: [Wien] Energy convergence with spin-orbit

[Islam, Md F]

Hi Prof. Marks,

Thanks for your comments. The calculation is with single impurity, so there 
is no issue with FM or AFM alignment.
I also chose the magnetization such that symmetry is not broken for in-plane or 
out-of-plane orientation. This particular
calculation is done with Fe impurity but I have done the same set of 
calculations with Mn impurity on the same surface
(Bi2Se3) without any problem. I agree that it doesn't guarantee a solution for 
all calculations but I just wanted to make
sure I have explored all possibilities before give up on it. Here are some 
energies around which SCF is oscillating.


:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79554107
:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79537485
:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79534902

:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79494494
:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79504374
:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79510509

:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79523383
:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79548735
:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79554582

This is going on for several weeks, so it doesn't look like converging slowly.

Thanks,
Fhokrul



From: wien-boun...@zeus.theochem.tuwien.ac.at 
[wien-boun...@zeus.theochem.tuwien.ac.at] On Behalf Of Laurence Marks 
[laurence.ma...@gmail.com]
Sent: Friday, February 12, 2016 7:49 AM
To: A Mailing list for WIEN2k users
Subject: Re: [Wien] Energy convergence with spin-orbit

I have only done a few -so calculations, but from the list they seem to lead to 
problems. A few suggestions:

a) Check for user error when you setup the in plane calculation. Have you used 
the right symmetry reduction? This seems to be the most common issue.

b) Check that you are using the patches for issues in initso that have been 
posted.

c) Look if your calculation is oscillating or just veery slowly converging. 
You can do a grep on some of the moments such as :ORB, :POM, :SPI in case.scf. 
I have seen these only slowly converge as the occupied states change.

d) Think carefully about the physics of your problem. Some time ago I did a 
test LDA+U calculation for a Ni (metal) surface. It did not converge, which was 
OK. In Wien2k you are solving a fixed-point problem, and always assuming that 
such a solution exists. I do not believe that anything  guaruntees that a 
solution exists in all cases. ( This is different from doing a variational 
calculation.) Maybe with your distances your impurities want to have an AFM 
ordering in-plane, but you are preventing this.

These are just guesses, without more information it is hard to know.

---
Professor Laurence Marks
Department of Materials Science and Engineering
Northwestern University
http://www.numis.northwestern.edu
Corrosion in 4D http://MURI4D.numis.northwestern.edu
Co-Editor, Acta Cryst A
"Research is to see what everybody else has seen, and to think what nobody else 
has thought"
Albert Szent-Gyorgi

On Feb 12, 2016 06:19, "Islam, Md F" <isl...@uta.edu<mailto:isl...@uta.edu>> 
wrote:
Dear Wien2k users,

I am trying to do a magnetic impurity calculation on a surface including 
spin-orbit coupling.
I set the energy and charge convergence to 0.1 and 0.001, respectively. The 
calculation is
well converged if magnetization is in-plane but for out of plane magnetization 
it is not converging.
While charge satisfies convergence criterion very well, the energy keeps 
oscillating near the 3rd/4th
decimal places.. I tried with different mixing (MSR1: 0.05 to 0.01) but nothing 
seems to be working
for about 2 months. For larger mixing, job crashes with QTL-B error. I would 
appreciate if you can
give me some suggestion to resolve the problem.

   I am also wondering if anyone has done any anisotropy calculation with both 
self-consistent and
non self-consistent methods and whether they give the same anisotropy energy.


Thanks,
Fhokrul

___
Wien mailing list
Wien@zeus.theochem.tuwien.ac.at<mailto: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
___
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

Re: [Wien] Energy convergence with spin-orbit

[Laurence Marks]

The energy is not always a good measure. Do, for instance if the impurity
is atom 1,
 grep :POM001DN *scf

Is this oscillating or slowly changing? Check the other terms I suggested.

N.B., remember that you have periodic continuation in the surface plane, so
unless you have a large supercell you are imposing FM ordering.

On Fri, Feb 12, 2016 at 9:03 AM, Islam, Md F <isl...@uta.edu> wrote:

> Hi Prof. Marks,
>
> Thanks for your comments. The calculation is with single impurity, so
> there is no issue with FM or AFM alignment.
> I also chose the magnetization such that symmetry is not broken for
> in-plane or out-of-plane orientation. This particular
> calculation is done with Fe impurity but I have done the same set of
> calculations with Mn impurity on the same surface
> (Bi2Se3) without any problem. I agree that it doesn't guarantee a solution
> for all calculations but I just wanted to make
> sure I have explored all possibilities before give up on it. Here are some
> energies around which SCF is oscillating.
>
>
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79554107
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79537485
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79534902
>
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79494494
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79504374
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79510509
>
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79523383
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79548735
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79554582
>
> This is going on for several weeks, so it doesn't look like converging
> slowly.
>
> Thanks,
> Fhokrul
>
>
> 
> From: wien-boun...@zeus.theochem.tuwien.ac.at [
> wien-boun...@zeus.theochem.tuwien.ac.at] On Behalf Of Laurence Marks [
> laurence.ma...@gmail.com]
> Sent: Friday, February 12, 2016 7:49 AM
> To: A Mailing list for WIEN2k users
> Subject: Re: [Wien] Energy convergence with spin-orbit
>
> I have only done a few -so calculations, but from the list they seem to
> lead to problems. A few suggestions:
>
> a) Check for user error when you setup the in plane calculation. Have you
> used the right symmetry reduction? This seems to be the most common issue.
>
> b) Check that you are using the patches for issues in initso that have
> been posted.
>
> c) Look if your calculation is oscillating or just veery slowly
> converging. You can do a grep on some of the moments such as :ORB, :POM,
> :SPI in case.scf. I have seen these only slowly converge as the occupied
> states change.
>
> d) Think carefully about the physics of your problem. Some time ago I did
> a test LDA+U calculation for a Ni (metal) surface. It did not converge,
> which was OK. In Wien2k you are solving a fixed-point problem, and always
> assuming that such a solution exists. I do not believe that anything
> guaruntees that a solution exists in all cases. ( This is different from
> doing a variational calculation.) Maybe with your distances your impurities
> want to have an AFM ordering in-plane, but you are preventing this.
>
> These are just guesses, without more information it is hard to know.
>
> ---
> Professor Laurence Marks
> Department of Materials Science and Engineering
> Northwestern University
> http://www.numis.northwestern.edu
> Corrosion in 4D http://MURI4D.numis.northwestern.edu
> Co-Editor, Acta Cryst A
> "Research is to see what everybody else has seen, and to think what nobody
> else has thought"
> Albert Szent-Gyorgi
>
> On Feb 12, 2016 06:19, "Islam, Md F" <isl...@uta.edu<mailto:isl...@uta.edu>>
> wrote:
> Dear Wien2k users,
>
> I am trying to do a magnetic impurity calculation on a surface
> including spin-orbit coupling.
> I set the energy and charge convergence to 0.1 and 0.001,
> respectively. The calculation is
> well converged if magnetization is in-plane but for out of plane
> magnetization it is not converging.
> While charge satisfies convergence criterion very well, the energy keeps
> oscillating near the 3rd/4th
> decimal places.. I tried with different mixing (MSR1: 0.05 to 0.01) but
> nothing seems to be working
> for about 2 months. For larger mixing, job crashes with QTL-B error. I
> would appreciate if you can
> give me some suggestion to resolve the problem.
>
>I am also wondering if anyone has done any anisotropy calculation with
> both self-consistent and
> non self-consistent methods and whether they give the same anisotropy
> energy.
>
>
> Thanks,
> Fhokrul
>
> _

Re: [Wien] Energy convergence with spin-orbit

[Islam, Md F]

I have checked :ORB, :POM, :SPI as you suggested but case.scf doesn't seem
to have any of those listed in it (doesn't return anything). I think :ORB works 
only if the calculation is done with LDA+U but my calculation is without U. 

Thanks,
Fhokrul



From: wien-boun...@zeus.theochem.tuwien.ac.at 
[wien-boun...@zeus.theochem.tuwien.ac.at] On Behalf Of Laurence Marks 
[l-ma...@northwestern.edu]
Sent: Friday, February 12, 2016 9:14 AM
To: A Mailing list for WIEN2k users
Subject: Re: [Wien] Energy convergence with spin-orbit

The energy is not always a good measure. Do, for instance if the impurity is 
atom 1,
 grep :POM001DN *scf

Is this oscillating or slowly changing? Check the other terms I suggested.

N.B., remember that you have periodic continuation in the surface plane, so 
unless you have a large supercell you are imposing FM ordering.

On Fri, Feb 12, 2016 at 9:03 AM, Islam, Md F 
<isl...@uta.edu<mailto:isl...@uta.edu>> wrote:
Hi Prof. Marks,

Thanks for your comments. The calculation is with single impurity, so there 
is no issue with FM or AFM alignment.
I also chose the magnetization such that symmetry is not broken for in-plane or 
out-of-plane orientation. This particular
calculation is done with Fe impurity but I have done the same set of 
calculations with Mn impurity on the same surface
(Bi2Se3) without any problem. I agree that it doesn't guarantee a solution for 
all calculations but I just wanted to make
sure I have explored all possibilities before give up on it. Here are some 
energies around which SCF is oscillating.


:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79554107
:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79537485
:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79534902

:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79494494
:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79504374
:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79510509

:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79523383
:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79548735
:ENE  : ** TOTAL ENERGY IN Ry = -2381169.79554582

This is going on for several weeks, so it doesn't look like converging slowly.

Thanks,
Fhokrul



From: 
wien-boun...@zeus.theochem.tuwien.ac.at<mailto:wien-boun...@zeus.theochem.tuwien.ac.at>
 
[wien-boun...@zeus.theochem.tuwien.ac.at<mailto:wien-boun...@zeus.theochem.tuwien.ac.at>]
 On Behalf Of Laurence Marks 
[laurence.ma...@gmail.com<mailto:laurence.ma...@gmail.com>]
Sent: Friday, February 12, 2016 7:49 AM
To: A Mailing list for WIEN2k users
Subject: Re: [Wien] Energy convergence with spin-orbit

I have only done a few -so calculations, but from the list they seem to lead to 
problems. A few suggestions:

a) Check for user error when you setup the in plane calculation. Have you used 
the right symmetry reduction? This seems to be the most common issue.

b) Check that you are using the patches for issues in initso that have been 
posted.

c) Look if your calculation is oscillating or just veery slowly converging. 
You can do a grep on some of the moments such as :ORB, :POM, :SPI in case.scf. 
I have seen these only slowly converge as the occupied states change.

d) Think carefully about the physics of your problem. Some time ago I did a 
test LDA+U calculation for a Ni (metal) surface. It did not converge, which was 
OK. In Wien2k you are solving a fixed-point problem, and always assuming that 
such a solution exists. I do not believe that anything  guaruntees that a 
solution exists in all cases. ( This is different from doing a variational 
calculation.) Maybe with your distances your impurities want to have an AFM 
ordering in-plane, but you are preventing this.

These are just guesses, without more information it is hard to know.

---
Professor Laurence Marks
Department of Materials Science and Engineering
Northwestern University
http://www.numis.northwestern.edu
Corrosion in 4D http://MURI4D.numis.northwestern.edu
Co-Editor, Acta Cryst A
"Research is to see what everybody else has seen, and to think what nobody else 
has thought"
Albert Szent-Gyorgi

On Feb 12, 2016 06:19, "Islam, Md F" 
<isl...@uta.edu<mailto:isl...@uta.edu><mailto:isl...@uta.edu<mailto:isl...@uta.edu>>>
 wrote:
Dear Wien2k users,

I am trying to do a magnetic impurity calculation on a surface including 
spin-orbit coupling.
I set the energy and charge convergence to 0.1 and 0.001, respectively. The 
calculation is
well converged if magnetization is in-plane but for out of plane magnetization 
it is not converging.
While charge satisfies convergence criterion very well, the energy keeps 
oscillating near the 3rd/4th
decimal places.. I tried with different mixing (MSR1: 0.05 to 0.01) but nothing 
seems to be working
for about 2 months. For larger mixing, job crashes with QTL-

Re: [Wien] Energy convergence with spin-orbit

[Laurence Marks]

Yes. I checked and I am 99% certain that you can add -dm which will then
give this information. You may be able to see something similar by doing a
grep on :QTL001 (or another number), although that is harder to decipher.

I have noticed that with -so swinging the occupation from one d-orbital to
another when they have very similar energies can be slow. I can rationalize
this with an argument that the mixer is a linear model and a rotation of
the d-orbital occupancies is non-linear, but this is waving my hands.

Quite often it looks like the iterations are not doing anything if you just
look at :DIS or :ENE, but a deeper look at other metrics (:CTO sometimes or
the magnetism) shows that it is slowly changing.

Someone else may have a better idea. This is all I can suggest based upon
partial knowledge.

On Fri, Feb 12, 2016 at 9:41 AM, Islam, Md F <isl...@uta.edu> wrote:

> I have checked :ORB, :POM, :SPI as you suggested but case.scf doesn't seem
> to have any of those listed in it (doesn't return anything). I think :ORB
> works
> only if the calculation is done with LDA+U but my calculation is without U.
>
> Thanks,
> Fhokrul
>
>
> 
> From: wien-boun...@zeus.theochem.tuwien.ac.at [
> wien-boun...@zeus.theochem.tuwien.ac.at] On Behalf Of Laurence Marks [
> l-ma...@northwestern.edu]
> Sent: Friday, February 12, 2016 9:14 AM
> To: A Mailing list for WIEN2k users
> Subject: Re: [Wien] Energy convergence with spin-orbit
>
> The energy is not always a good measure. Do, for instance if the impurity
> is atom 1,
>  grep :POM001DN *scf
>
> Is this oscillating or slowly changing? Check the other terms I suggested.
>
> N.B., remember that you have periodic continuation in the surface plane,
> so unless you have a large supercell you are imposing FM ordering.
>
> On Fri, Feb 12, 2016 at 9:03 AM, Islam, Md F <isl...@uta.edu isl...@uta.edu>> wrote:
> Hi Prof. Marks,
>
> Thanks for your comments. The calculation is with single impurity, so
> there is no issue with FM or AFM alignment.
> I also chose the magnetization such that symmetry is not broken for
> in-plane or out-of-plane orientation. This particular
> calculation is done with Fe impurity but I have done the same set of
> calculations with Mn impurity on the same surface
> (Bi2Se3) without any problem. I agree that it doesn't guarantee a solution
> for all calculations but I just wanted to make
> sure I have explored all possibilities before give up on it. Here are some
> energies around which SCF is oscillating.
>
>
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79554107
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79537485
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79534902
>
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79494494
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79504374
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79510509
>
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79523383
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79548735
> :ENE  : ** TOTAL ENERGY IN Ry = -2381169.79554582
>
> This is going on for several weeks, so it doesn't look like converging
> slowly.
>
> Thanks,
> Fhokrul
>
>
> 
> From: wien-boun...@zeus.theochem.tuwien.ac.at wien-boun...@zeus.theochem.tuwien.ac.at> [
> wien-boun...@zeus.theochem.tuwien.ac.at wien-boun...@zeus.theochem.tuwien.ac.at>] On Behalf Of Laurence Marks [
> laurence.ma...@gmail.com<mailto:laurence.ma...@gmail.com>]
> Sent: Friday, February 12, 2016 7:49 AM
> To: A Mailing list for WIEN2k users
> Subject: Re: [Wien] Energy convergence with spin-orbit
>
> I have only done a few -so calculations, but from the list they seem to
> lead to problems. A few suggestions:
>
> a) Check for user error when you setup the in plane calculation. Have you
> used the right symmetry reduction? This seems to be the most common issue.
>
> b) Check that you are using the patches for issues in initso that have
> been posted.
>
> c) Look if your calculation is oscillating or just veery slowly
> converging. You can do a grep on some of the moments such as :ORB, :POM,
> :SPI in case.scf. I have seen these only slowly converge as the occupied
> states change.
>
> d) Think carefully about the physics of your problem. Some time ago I did
> a test LDA+U calculation for a Ni (metal) surface. It did not converge,
> which was OK. In Wien2k you are solving a fixed-point problem, and always
> assuming that such a solution exists. I do not believe that anything
> guaruntees that a solution exists in all cases. ( This is different from
> doing a variational calculation.) Maybe with