Re: [Wien] calculation with lmbj potential

[fabien . tran]

Hi,

In principle, mBJ can not be applied to systems with vacuum or an 
interface (see section "Modified Becke-Johnson potential (mBJ) for band 
gaps" in the user's guide). An alternative is to use lmBJ as you did, 
but convergence was not possible. Another possibility is to use mBJ, but 
by fixing the value of grad(rho)/rho to some (maybe arbitrary) value 
(see explanations in the UG).


How bad is the convergence with PBE+nlvdw? Are you also trying to relax 
the structure? Can you show the input files case.in0 and case.innlvdw?



On 26.07.2023 18:35, Burhan Ahmed wrote:

I check the possible ways of converging my system.

* The system converges with PBE+SO
* The system converges with TB-mBJ+SO

But

* The system doesn’t converges lmBJ+SO
* The system doesn’t converges with PBE+nlvdw

It seems whenever I try to incorporate the van der waals interactions,
the system becomes unstable.

Karnel Type I choose is 1 and the rest default parameters are used.

I am still searching for the possible solution. Hoping for suggestions
from the experts.

Regards

Burhan Ahmed

Research Scholar, AUS

From: Laurence Marks
Sent: Monday, July 24, 2023 3:44 PM
To: A Mailing list for WIEN2k users
Subject: Re: [Wien] calculation with lmbj potential

You should look at the BVS, i.e. "grep Bond *tnn" and compare it to
what you have for bulk. You will see that you surface has bad values,
so will be unstable. You need to do a lot more thinking and analysis
(weeks) to find a chemically reasonable surface.

---
Professor Laurence Marks (Laurie)
Department of Materials Science and Engineering
Northwestern University
www.numis.northwestern.edu [1]
"Research is to see what everybody else has seen, and to think what
nobody else has thought" Albert Szent-Györgyi

On Mon, Jul 24, 2023, 05:28 Peter Blaha 
wrote:

Yes, R cells are first converted into H (making it already 3 times
larger than the primitive R cell).

The "accepting repeat atoms at z=0" makes the non-stoichiometry. It
leads to 2 identical surfaces and inversion symmetry (cheaper calc),
but
non-stoiciometry.
It is not always clear what the best model for a surface is. Yours is
Te
terminated, but maybe another termination is more favorable (Bi, or
maybe another Te-layer ?)

Am 24.07.2023 um 10:51 schrieb Burhan Ahmed:

The 6ql slab is created with 1x1x2 supercell with vacuum in Z

direction

and accepting repeat atoms at z=0 (using x supercell program). At

first

the x supercell convert rhombohedral cell into hexagonal and then

from

hexagonal cell I have created the 1x1x2 supercell and then I took

the

structure suggested by sgroup.

Yes sorry for that I have used MSR1a method for the relaxation.

The force convergence is set to 1 Ry.

On Mon, 24 Jul, 2023, 2:06 pm Peter Blaha, 


<mailto:peter.bl...@tuwien.ac.at>> wrote:

I'm not sure how you get a non-stoichimetric cell with a

multiplicative

number of unit cells, unless you said you want to repeat the

atom at

z=0.
Of course, without this extra layer, you may not have inversion

and get

2 different surface terminations in one calculation. This is the

usual

problem of slabs.
Also symmetry will be reduced and "multiplicity" errors occur.

But

both,
nn and sgroup create new struct files where this has been

corrected.


If possible, I would also go to a smaller number of layers.


Am 24.07.2023 um 10:13 schrieb Burhan Ahmed:
 > That is very true. I made the slab using the tutorial

available

in the
 > wien2k user manual by executing x-supercell program. What I

found is

 > that the 6ql supercell consist of an extra Te atom. But

whenever

I try
 > to remove this atom I got multiplicity error. Sir, what is

best

possible
 > way of making slab with vacuum or surfaces??
 >
 > Thanks
 >
 > Regards
 >
 > Burhan Ahmed
 >
 > *Research Scholar, AUS *
 >
 > *From: *Laurence Marks <mailto:laurence.ma...@gmail.com
<mailto:laurence.ma...@gmail.com>>
 > *Sent: *Saturday, July 22, 2023 6:30 PM
 > *To: *A Mailing list for WIEN2k users
 > <mailto:wien@zeus.theochem.tuwien.ac.at
<mailto:wien@zeus.theochem.tuwien.ac.at>>
 > *Subject: *Re: [Wien] calculation with lmbj potential
 >
 > Others will probably give you suggestions about converging

mBJ. Some

 > deeper comments.
 >
 > The most common reason that calculations behave badly is user

error.

 > Sometimes this is doing the initialization wrong, often it is
creating
 > an inappropriate model. Just because one can use x supercell

or

related
 > codes to create atomic positions does /not/ make them

sensible.

 >
 > Your slab has a composition Bi12 Te19, i.e. it is Te rich.

You

therefore

Re: [Wien] calculation with lmbj potential

[Burhan Ahmed]

I check the possible ways of converging my system. The system converges with PBE+SOThe system converges with TB-mBJ+SO But The system doesn’t converges lmBJ+SOThe system doesn’t converges with PBE+nlvdwIt seems whenever I try to incorporate the van der waals interactions, the system becomes unstable. Karnel Type I choose is 1 and the rest default parameters are used.  I am still searching for the possible solution. Hoping for suggestions from the experts.   RegardsBurhan AhmedResearch Scholar, AUS  From: Laurence MarksSent: Monday, July 24, 2023 3:44 PMTo: A Mailing list for WIEN2k usersSubject: Re: [Wien] calculation with lmbj potential You should look at the BVS, i.e. "grep Bond *tnn" and compare it to what you have for bulk. You will see that you surface has bad values, so will be unstable. You need to do a lot more thinking and analysis (weeks) to find a chemically reasonable surface.---Professor Laurence Marks (Laurie)Department of Materials Science and EngineeringNorthwestern Universitywww.numis.northwestern.edu"Research is to see what everybody else has seen, and to think what nobody else has thought" Albert Szent-Györgyi On Mon, Jul 24, 2023, 05:28 Peter Blaha <peter.bl...@tuwien.ac.at> wrote:Yes, R cells are first converted into H (making it already 3 times larger than the primitive R cell).The "accepting repeat atoms at z=0" makes the non-stoichiometry. It leads to 2 identical surfaces and inversion symmetry (cheaper calc), but non-stoiciometry.It is not always clear what the best model for a surface is. Yours is Te terminated, but maybe another termination is more favorable (Bi, or maybe another Te-layer ?)Am 24.07.2023 um 10:51 schrieb Burhan Ahmed:> The 6ql slab is created with 1x1x2 supercell with vacuum in Z direction > and accepting repeat atoms at z=0 (using x supercell program). At first > the x supercell convert rhombohedral cell into hexagonal and then from > hexagonal cell I have created the 1x1x2 supercell and then I took the > structure suggested by sgroup.> > Yes sorry for that I have used MSR1a method for the relaxation.> > The force convergence is set to 1 Ry.> > On Mon, 24 Jul, 2023, 2:06 pm Peter Blaha, <peter.bl...@tuwien.ac.at > peter.bl...@tuwien.ac.at>> wrote:> >     I'm not sure how you get a non-stoichimetric cell with a multiplicative>     number of unit cells, unless you said you want to repeat the atom at>     z=0.>     Of course, without this extra layer, you may not have inversion and get>     2 different surface terminations in one calculation. This is the usual>     problem of slabs.>     Also symmetry will be reduced and "multiplicity" errors occur. But>     both,>     nn and sgroup create new struct files where this has been corrected.> >     If possible, I would also go to a smaller number of layers.> > >     Am 24.07.2023 um 10:13 schrieb Burhan Ahmed:>      > That is very true. I made the slab using the tutorial available>     in the>      > wien2k user manual by executing x-supercell program. What I found is>      > that the 6ql supercell consist of an extra Te atom. But whenever>     I try>      > to remove this atom I got multiplicity error. Sir, what is best>     possible>      > way of making slab with vacuum or surfaces??>      >>      > Thanks>      >>      > Regards>      >>      > Burhan Ahmed>      >>      > *Research Scholar, AUS *>      >>      > *From: *Laurence Marks laurence.ma...@gmail.com>     laurence.ma...@gmail.com>>>      > *Sent: *Saturday, July 22, 2023 6:30 PM>      > *To: *A Mailing list for WIEN2k users>      > wien@zeus.theochem.tuwien.ac.at>     wien@zeus.theochem.tuwien.ac.at>>>      > *Subject: *Re: [Wien] calculation with lmbj potential>      >>      > Others will probably give you suggestions about converging mBJ. Some>      > deeper comments.>      >>      > The most common reason that calculations behave badly is user error.>      > Sometimes this is doing the initialization wrong, often it is>     creating>      > an inappropriate model. Just because one can use x supercell or>     related>      > codes to create atomic positions does /not/ make them sensible.>      >>      > Your slab has a composition Bi12 Te19, i.e. it is Te rich. You>     therefore>      > have a reduced, n-type semiconductor with a small gap (about 0.1eV)>      > which will behave badly. If you look at your BVS you will see>     that atom>      > Te7 is severely underco-ordinated (unstable).>      >>      > I doubt that your slab will ever converge to anything which>     should be>      > published. Surfaces are not simple, it is unfortunately too easy>     to

Re: [Wien] calculation with lmbj potential

[Laurence Marks]

You should look at the BVS, i.e. "grep Bond *tnn" and compare it to what
you have for bulk. You will see that you surface has bad values, so will be
unstable. You need to do a lot more thinking and analysis (weeks) to find a
chemically reasonable surface.

---
Professor Laurence Marks (Laurie)
Department of Materials Science and Engineering
Northwestern University
www.numis.northwestern.edu
"Research is to see what everybody else has seen, and to think what nobody
else has thought" Albert Szent-Györgyi

On Mon, Jul 24, 2023, 05:28 Peter Blaha  wrote:

> Yes, R cells are first converted into H (making it already 3 times
> larger than the primitive R cell).
>
> The "accepting repeat atoms at z=0" makes the non-stoichiometry. It
> leads to 2 identical surfaces and inversion symmetry (cheaper calc), but
> non-stoiciometry.
> It is not always clear what the best model for a surface is. Yours is Te
> terminated, but maybe another termination is more favorable (Bi, or
> maybe another Te-layer ?)
>
> Am 24.07.2023 um 10:51 schrieb Burhan Ahmed:
> > The 6ql slab is created with 1x1x2 supercell with vacuum in Z direction
> > and accepting repeat atoms at z=0 (using x supercell program). At first
> > the x supercell convert rhombohedral cell into hexagonal and then from
> > hexagonal cell I have created the 1x1x2 supercell and then I took the
> > structure suggested by sgroup.
> >
> > Yes sorry for that I have used MSR1a method for the relaxation.
> >
> > The force convergence is set to 1 Ry.
> >
> > On Mon, 24 Jul, 2023, 2:06 pm Peter Blaha,  > <mailto:peter.bl...@tuwien.ac.at>> wrote:
> >
> > I'm not sure how you get a non-stoichimetric cell with a
> multiplicative
> > number of unit cells, unless you said you want to repeat the atom at
> > z=0.
> > Of course, without this extra layer, you may not have inversion and
> get
> > 2 different surface terminations in one calculation. This is the
> usual
> > problem of slabs.
> > Also symmetry will be reduced and "multiplicity" errors occur. But
> > both,
> > nn and sgroup create new struct files where this has been corrected.
> >
> > If possible, I would also go to a smaller number of layers.
> >
> >
> > Am 24.07.2023 um 10:13 schrieb Burhan Ahmed:
> >  > That is very true. I made the slab using the tutorial available
> > in the
> >  > wien2k user manual by executing x-supercell program. What I found
> is
> >  > that the 6ql supercell consist of an extra Te atom. But whenever
> > I try
> >  > to remove this atom I got multiplicity error. Sir, what is best
> > possible
> >  > way of making slab with vacuum or surfaces??
> >  >
> >  > Thanks
> >  >
> >  > Regards
> >  >
> >  > Burhan Ahmed
> >  >
> >      > *Research Scholar, AUS *
> >  >
> >  > *From: *Laurence Marks <mailto:laurence.ma...@gmail.com
> > <mailto:laurence.ma...@gmail.com>>
> >  > *Sent: *Saturday, July 22, 2023 6:30 PM
> >  > *To: *A Mailing list for WIEN2k users
> >  > <mailto:wien@zeus.theochem.tuwien.ac.at
> > <mailto:wien@zeus.theochem.tuwien.ac.at>>
> >  > *Subject: *Re: [Wien] calculation with lmbj potential
> >  >
> >  > Others will probably give you suggestions about converging mBJ.
> Some
> >  > deeper comments.
> >  >
> >  > The most common reason that calculations behave badly is user
> error.
> >  > Sometimes this is doing the initialization wrong, often it is
> > creating
> >  > an inappropriate model. Just because one can use x supercell or
> > related
> >  > codes to create atomic positions does /not/ make them sensible.
> >  >
> >  > Your slab has a composition Bi12 Te19, i.e. it is Te rich. You
> > therefore
> >  > have a reduced, n-type semiconductor with a small gap (about
> 0.1eV)
> >  > which will behave badly. If you look at your BVS you will see
> > that atom
> >  > Te7 is severely underco-ordinated (unstable).
> >  >
> >  > I doubt that your slab will ever converge to anything which
> > should be
> >  > published. Surfaces are not simple, it is unfortunately too easy
> > to set
> >  > something up and get it wrong. (There are many papers in the
> > literature
> >  > of dubi

Re: [Wien] calculation with lmbj potential

[Peter Blaha]

Yes, R cells are first converted into H (making it already 3 times 
larger than the primitive R cell).


The "accepting repeat atoms at z=0" makes the non-stoichiometry. It 
leads to 2 identical surfaces and inversion symmetry (cheaper calc), but 
non-stoiciometry.
It is not always clear what the best model for a surface is. Yours is Te 
terminated, but maybe another termination is more favorable (Bi, or 
maybe another Te-layer ?)


Am 24.07.2023 um 10:51 schrieb Burhan Ahmed:
The 6ql slab is created with 1x1x2 supercell with vacuum in Z direction 
and accepting repeat atoms at z=0 (using x supercell program). At first 
the x supercell convert rhombohedral cell into hexagonal and then from 
hexagonal cell I have created the 1x1x2 supercell and then I took the 
structure suggested by sgroup.


Yes sorry for that I have used MSR1a method for the relaxation.

The force convergence is set to 1 Ry.

On Mon, 24 Jul, 2023, 2:06 pm Peter Blaha, <mailto:peter.bl...@tuwien.ac.at>> wrote:


I'm not sure how you get a non-stoichimetric cell with a multiplicative
number of unit cells, unless you said you want to repeat the atom at
z=0.
Of course, without this extra layer, you may not have inversion and get
2 different surface terminations in one calculation. This is the usual
problem of slabs.
Also symmetry will be reduced and "multiplicity" errors occur. But
both,
nn and sgroup create new struct files where this has been corrected.

If possible, I would also go to a smaller number of layers.


Am 24.07.2023 um 10:13 schrieb Burhan Ahmed:
 > That is very true. I made the slab using the tutorial available
in the
 > wien2k user manual by executing x-supercell program. What I found is
 > that the 6ql supercell consist of an extra Te atom. But whenever
I try
 > to remove this atom I got multiplicity error. Sir, what is best
possible
 > way of making slab with vacuum or surfaces??
 >
 > Thanks
 >
 > Regards
 >
 > Burhan Ahmed
 >
 > *Research Scholar, AUS *
 >
 > *From: *Laurence Marks <mailto:laurence.ma...@gmail.com
<mailto:laurence.ma...@gmail.com>>
 > *Sent: *Saturday, July 22, 2023 6:30 PM
 > *To: *A Mailing list for WIEN2k users
 > <mailto:wien@zeus.theochem.tuwien.ac.at
<mailto:wien@zeus.theochem.tuwien.ac.at>>
 > *Subject: *Re: [Wien] calculation with lmbj potential
 >
 > Others will probably give you suggestions about converging mBJ. Some
 > deeper comments.
 >
 > The most common reason that calculations behave badly is user error.
 > Sometimes this is doing the initialization wrong, often it is
creating
 > an inappropriate model. Just because one can use x supercell or
related
 > codes to create atomic positions does /not/ make them sensible.
 >
 > Your slab has a composition Bi12 Te19, i.e. it is Te rich. You
therefore
 > have a reduced, n-type semiconductor with a small gap (about 0.1eV)
 > which will behave badly. If you look at your BVS you will see
that atom
 > Te7 is severely underco-ordinated (unstable).
 >
 > I doubt that your slab will ever converge to anything which
should be
 > published. Surfaces are not simple, it is unfortunately too easy
to set
 > something up and get it wrong. (There are many papers in the
literature
 > of dubious merit.)
 >
 > On Fri, Jul 21, 2023 at 10:47 PM Burhan Ahmed
mailto:burhan.ah...@aus.ac.in>
 > <mailto:burhan.ah...@aus.ac.in <mailto:burhan.ah...@aus.ac.in>>>
wrote:
 >
 >     Dear experts, I am doing an scf calculation taking lmbj
potential.
 >     My system is a slab of 6ql with a vacuum of 40 ang along c-axix.
 >     Whenever I try to run the scf calculation including SOC,
after 999
 >     iteration the scf is still not converged. I have analyzed the scf
 >     file, it shows the fluctuating nature. At first the selected
Rmt was
 >     2.5 then I reduces it to 2.34 and again convergence failed
after 999
 >     cycle. I have included -hdlo and -lvns 8 switch in init_lapw
as it
 >     shows heavy atom. I am attaching the case.struct file. Hoping for
 >     any suggestion/solution.
 >
 >     Regards
 >
 >     Burhan Ahmed
 >
 >     *Research Scholar, AUS *
 >
 >     ___
 >     Wien mailing list
 > Wien@zeus.theochem.tuwien.ac.at
<mailto:Wien@zeus.theochem.tuwien.ac.at>
<mailto:Wien@zeus.theochem.tuwien.ac.at
<mailto:Wien@zeus.theochem.tuwien.ac.at>>
 > http://zeus.theochem.tuwien.ac.

Re: [Wien] calculation with lmbj potential

[Burhan Ahmed]

The 6ql slab is created with 1x1x2 supercell with vacuum in Z direction and
accepting repeat atoms at z=0 (using x supercell program). At first the x
supercell convert rhombohedral cell into hexagonal and then from hexagonal
cell I have created the 1x1x2 supercell and then I took the structure
suggested by sgroup.

Yes sorry for that I have used MSR1a method for the relaxation.

The force convergence is set to 1 Ry.

On Mon, 24 Jul, 2023, 2:06 pm Peter Blaha,  wrote:

> I'm not sure how you get a non-stoichimetric cell with a multiplicative
> number of unit cells, unless you said you want to repeat the atom at z=0.
> Of course, without this extra layer, you may not have inversion and get
> 2 different surface terminations in one calculation. This is the usual
> problem of slabs.
> Also symmetry will be reduced and "multiplicity" errors occur. But both,
> nn and sgroup create new struct files where this has been corrected.
>
> If possible, I would also go to a smaller number of layers.
>
>
> Am 24.07.2023 um 10:13 schrieb Burhan Ahmed:
> > That is very true. I made the slab using the tutorial available in the
> > wien2k user manual by executing x-supercell program. What I found is
> > that the 6ql supercell consist of an extra Te atom. But whenever I try
> > to remove this atom I got multiplicity error. Sir, what is best possible
> > way of making slab with vacuum or surfaces??
> >
> > Thanks
> >
> > Regards
> >
> > Burhan Ahmed
> >
> > *Research Scholar, AUS *
> >
> > *From: *Laurence Marks <mailto:laurence.ma...@gmail.com>
> > *Sent: *Saturday, July 22, 2023 6:30 PM
> > *To: *A Mailing list for WIEN2k users
> > <mailto:wien@zeus.theochem.tuwien.ac.at>
> > *Subject: *Re: [Wien] calculation with lmbj potential
> >
> > Others will probably give you suggestions about converging mBJ. Some
> > deeper comments.
> >
> > The most common reason that calculations behave badly is user error.
> > Sometimes this is doing the initialization wrong, often it is creating
> > an inappropriate model. Just because one can use x supercell or related
> > codes to create atomic positions does /not/ make them sensible.
> >
> > Your slab has a composition Bi12 Te19, i.e. it is Te rich. You therefore
> > have a reduced, n-type semiconductor with a small gap (about 0.1eV)
> > which will behave badly. If you look at your BVS you will see that atom
> > Te7 is severely underco-ordinated (unstable).
> >
> > I doubt that your slab will ever converge to anything which should be
> > published. Surfaces are not simple, it is unfortunately too easy to set
> > something up and get it wrong. (There are many papers in the literature
> > of dubious merit.)
> >
> > On Fri, Jul 21, 2023 at 10:47 PM Burhan Ahmed  > <mailto:burhan.ah...@aus.ac.in>> wrote:
> >
> > Dear experts, I am doing an scf calculation taking lmbj potential.
> > My system is a slab of 6ql with a vacuum of 40 ang along c-axix.
> > Whenever I try to run the scf calculation including SOC, after 999
> > iteration the scf is still not converged. I have analyzed the scf
> > file, it shows the fluctuating nature. At first the selected Rmt was
> > 2.5 then I reduces it to 2.34 and again convergence failed after 999
> > cycle. I have included -hdlo and -lvns 8 switch in init_lapw as it
> > shows heavy atom. I am attaching the case.struct file. Hoping for
> > any suggestion/solution.
> >
> > Regards
> >
> > Burhan Ahmed
> >
> > *Research Scholar, AUS *
> >
> > ___
> > Wien mailing list
> > Wien@zeus.theochem.tuwien.ac.at  Wien@zeus.theochem.tuwien.ac.at>
> > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> > <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 <
> http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html>
> >
> >
> > --
> >
> > Professor Laurence Marks (Laurie)
> > Department of Materials Science and Engineering
> > Northwestern University
> > www.numis.northwestern.edu <http://www.numis.northwestern.edu>
> > "Research is to see what everybody else has seen, and to think what
> > nobody else has thought", Albert Szent-Györgyi
> >
> >
> > ___
> > Wien mailing list
> > Wien@zeus.theochem.tuwien.ac.at
> > http

Re: [Wien] calculation with lmbj potential

[Peter Blaha]

sgroup does NOT add atoms. It will only split or combine positions (and 
eventually even reduce the unit cell). It must come from supercell.


PS: I hope you have used MSR1a method for relaxation ? MSR1 does not 
move atoms. Are the final forces all small ?



Am 24.07.2023 um 10:30 schrieb Burhan Ahmed:

Yes, I have the relax the atomic positions using msr1 method.

I my 6ql slab I have one extra Te atom that generates from sgroup. Is 
there any problem with that? Or it just created for symmetry purposes??


On Mon, 24 Jul, 2023, 1:56 pm Peter Blaha, > wrote:


Hi,
a) Your vacuum is very large, which can create problems and cost a lot
of cpu time. 20 Ang are more than enough.
b) As mentioned before: this is a doped SC, so most likely never a gap
 Is it intended ?

d) Did you relax the slab in PBE ? Just the bulk coordinates are
probably nonsense.
e) ALWAYS converge PBE first; save_lapw; and use this as start for mBJ.


Am 24.07.2023 um 10:04 schrieb Burhan Ahmed:
 > The results from the last 20 iterations (for lmbj calculation)
 >
 > Analysis of parameter:
 > :ENE :DIS :GAP
 > in bi2te3lmbj.scf (showing last 20 / 1 lines)
 >
 > --- ENE ---
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.11941651
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.11305258
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.10571778
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.09359536
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.08522256
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.07976176
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.07273400
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.06627481
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.05653160
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.05078986
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.04165986
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.02824173
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.02513981
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.01487899
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775750.00743085
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775749.99627176
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775749.99306205
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775749.98317825
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775749.97811958
 > :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -775749.96939695
 > --- DIS ---
 > :DIS  :  CHARGE DISTANCE      (  4.2460241 for atom    3 spin 1) 
     2.0217470
 > :DIS  :  CHARGE DISTANCE      (  4.2354501 for atom    3 spin 1) 
     2.0167505
 > :DIS  :  CHARGE DISTANCE      (  4.2249023 for atom    3 spin 1) 
     2.0117662
 > :DIS  :  CHARGE DISTANCE      (  4.2143828 for atom    3 spin 1) 
     2.0067942
 > :DIS  :  CHARGE DISTANCE      (  4.2038864 for atom    3 spin 1) 
     2.0018343
 > :DIS  :  CHARGE DISTANCE      (  4.1934158 for atom    3 spin 1) 
     1.9968867
 > :DIS  :  CHARGE DISTANCE      (  4.1829711 for atom    3 spin 1) 
     1.9919508
 > :DIS  :  CHARGE DISTANCE      (  4.1725544 for atom    3 spin 1) 
     1.9870274
 > :DIS  :  CHARGE DISTANCE      (  4.1621656 for atom    3 spin 1) 
     1.9821164
 > :DIS  :  CHARGE DISTANCE      (  4.1518018 for atom    3 spin 1) 
     1.9772172
 > :DIS  :  CHARGE DISTANCE      (  4.1414653 for atom    3 spin 1) 
     1.9723300
 > :DIS  :  CHARGE DISTANCE      (  4.1311557 for atom    3 spin 1) 
     1.9674546
 > :DIS  :  CHARGE DISTANCE      (  4.1208686 for atom    3 spin 1) 
     1.9625908
 > :DIS  :  CHARGE DISTANCE      (  4.1106085 for atom    3 spin 1) 
     1.9577392
 > :DIS  :  CHARGE DISTANCE      (  4.1003736 for atom    3 spin 1) 
     1.9528994
 > :DIS  :  CHARGE DISTANCE      (  4.0901658 for atom    3 spin 1) 
     1.9480720
 > :DIS  :  CHARGE DISTANCE      (  4.0799795 for atom    3 spin 1) 
     1.9432557
 > :DIS  :  CHARGE DISTANCE      (  4.0698205 for atom    3 spin 1) 
     1.9384514
 > :DIS  :  CHARGE DISTANCE      (  4.0596869 for atom    3 spin 1) 
     1.9336590

 > :DIS  :  CHARGE DISTANCE      (  4.0495779 for atom    3 spin 1)
 > 1.9288781
 > #
 > Yes I have converged the calculation using PBE-GGA and PBE-GGA with
 > SO. And in both the cases I have found the metallic nature of crystal
 > (using grep :GAP) and also in band structure the VB crosses the Fermi
 > level.
 > ##
 > No, I didn't use the previous electron and kinetic energy density for
 > the lmbj calculation. What I did is created a new directory 

Re: [Wien] calculation with lmbj potential

[Peter Blaha]

I'm not sure how you get a non-stoichimetric cell with a multiplicative 
number of unit cells, unless you said you want to repeat the atom at z=0.
Of course, without this extra layer, you may not have inversion and get 
2 different surface terminations in one calculation. This is the usual 
problem of slabs.
Also symmetry will be reduced and "multiplicity" errors occur. But both, 
nn and sgroup create new struct files where this has been corrected.


If possible, I would also go to a smaller number of layers.


Am 24.07.2023 um 10:13 schrieb Burhan Ahmed:
That is very true. I made the slab using the tutorial available in the 
wien2k user manual by executing x-supercell program. What I found is 
that the 6ql supercell consist of an extra Te atom. But whenever I try 
to remove this atom I got multiplicity error. Sir, what is best possible 
way of making slab with vacuum or surfaces??


Thanks

Regards

Burhan Ahmed

*Research Scholar, AUS *

*From: *Laurence Marks <mailto:laurence.ma...@gmail.com>
*Sent: *Saturday, July 22, 2023 6:30 PM
*To: *A Mailing list for WIEN2k users 
<mailto:wien@zeus.theochem.tuwien.ac.at>

*Subject: *Re: [Wien] calculation with lmbj potential

Others will probably give you suggestions about converging mBJ. Some 
deeper comments.


The most common reason that calculations behave badly is user error. 
Sometimes this is doing the initialization wrong, often it is creating 
an inappropriate model. Just because one can use x supercell or related 
codes to create atomic positions does /not/ make them sensible.


Your slab has a composition Bi12 Te19, i.e. it is Te rich. You therefore 
have a reduced, n-type semiconductor with a small gap (about 0.1eV) 
which will behave badly. If you look at your BVS you will see that atom 
Te7 is severely underco-ordinated (unstable).


I doubt that your slab will ever converge to anything which should be 
published. Surfaces are not simple, it is unfortunately too easy to set 
something up and get it wrong. (There are many papers in the literature 
of dubious merit.)


On Fri, Jul 21, 2023 at 10:47 PM Burhan Ahmed <mailto:burhan.ah...@aus.ac.in>> wrote:


Dear experts, I am doing an scf calculation taking lmbj potential.
My system is a slab of 6ql with a vacuum of 40 ang along c-axix.
Whenever I try to run the scf calculation including SOC, after 999
iteration the scf is still not converged. I have analyzed the scf
file, it shows the fluctuating nature. At first the selected Rmt was
2.5 then I reduces it to 2.34 and again convergence failed after 999
cycle. I have included -hdlo and -lvns 8 switch in init_lapw as it
shows heavy atom. I am attaching the case.struct file. Hoping for
any suggestion/solution.

Regards

Burhan Ahmed

*Research Scholar, AUS *

___
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
<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 
<http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html>


--

Professor Laurence Marks (Laurie)
Department of Materials Science and Engineering
Northwestern University
www.numis.northwestern.edu <http://www.numis.northwestern.edu>
"Research is to see what everybody else has seen, and to think what 
nobody else has thought", Albert Szent-Györgyi



___
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
Email: peter.bl...@tuwien.ac.atWIEN2k: http://www.wien2k.at
WWW:   http://www.imc.tuwien.ac.at
-
___
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] calculation with lmbj potential

[Burhan Ahmed]

Yes, I have the relax the atomic positions using msr1 method.

I my 6ql slab I have one extra Te atom that generates from sgroup. Is there
any problem with that? Or it just created for symmetry purposes??

On Mon, 24 Jul, 2023, 1:56 pm Peter Blaha,  wrote:

> Hi,
> a) Your vacuum is very large, which can create problems and cost a lot
> of cpu time. 20 Ang are more than enough.
> b) As mentioned before: this is a doped SC, so most likely never a gap
>  Is it intended ?
>
> d) Did you relax the slab in PBE ? Just the bulk coordinates are
> probably nonsense.
> e) ALWAYS converge PBE first; save_lapw; and use this as start for mBJ.
>
>
> Am 24.07.2023 um 10:04 schrieb Burhan Ahmed:
> > The results from the last 20 iterations (for lmbj calculation)
> >
> > Analysis of parameter:
> > :ENE :DIS :GAP
> > in bi2te3lmbj.scf (showing last 20 / 1 lines)
> >
> > --- ENE ---
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.11941651
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.11305258
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.10571778
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.09359536
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.08522256
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.07976176
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.07273400
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.06627481
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.05653160
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.05078986
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.04165986
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.02824173
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.02513981
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.01487899
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.00743085
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.99627176
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.99306205
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.98317825
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.97811958
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.96939695
> > --- DIS ---
> > :DIS  :  CHARGE DISTANCE  (  4.2460241 for atom3 spin 1)
> 2.0217470
> > :DIS  :  CHARGE DISTANCE  (  4.2354501 for atom3 spin 1)
> 2.0167505
> > :DIS  :  CHARGE DISTANCE  (  4.2249023 for atom3 spin 1)
> 2.0117662
> > :DIS  :  CHARGE DISTANCE  (  4.2143828 for atom3 spin 1)
> 2.0067942
> > :DIS  :  CHARGE DISTANCE  (  4.2038864 for atom3 spin 1)
> 2.0018343
> > :DIS  :  CHARGE DISTANCE  (  4.1934158 for atom3 spin 1)
> 1.9968867
> > :DIS  :  CHARGE DISTANCE  (  4.1829711 for atom3 spin 1)
> 1.9919508
> > :DIS  :  CHARGE DISTANCE  (  4.1725544 for atom3 spin 1)
> 1.9870274
> > :DIS  :  CHARGE DISTANCE  (  4.1621656 for atom3 spin 1)
> 1.9821164
> > :DIS  :  CHARGE DISTANCE  (  4.1518018 for atom3 spin 1)
> 1.9772172
> > :DIS  :  CHARGE DISTANCE  (  4.1414653 for atom3 spin 1)
> 1.9723300
> > :DIS  :  CHARGE DISTANCE  (  4.1311557 for atom3 spin 1)
> 1.9674546
> > :DIS  :  CHARGE DISTANCE  (  4.1208686 for atom3 spin 1)
> 1.9625908
> > :DIS  :  CHARGE DISTANCE  (  4.1106085 for atom3 spin 1)
> 1.9577392
> > :DIS  :  CHARGE DISTANCE  (  4.1003736 for atom3 spin 1)
> 1.9528994
> > :DIS  :  CHARGE DISTANCE  (  4.0901658 for atom3 spin 1)
> 1.9480720
> > :DIS  :  CHARGE DISTANCE  (  4.0799795 for atom3 spin 1)
> 1.9432557
> > :DIS  :  CHARGE DISTANCE  (  4.0698205 for atom3 spin 1)
> 1.9384514
> > :DIS  :  CHARGE DISTANCE  (  4.0596869 for atom3 spin 1)
> 1.9336590
> > :DIS  :  CHARGE DISTANCE  (  4.0495779 for atom3 spin 1)
> > 1.9288781
> > #
> > Yes I have converged the calculation using PBE-GGA and PBE-GGA with
> > SO. And in both the cases I have found the metallic nature of crystal
> > (using grep :GAP) and also in band structure the VB crosses the Fermi
> > level.
> > ##
> > No, I didn't use the previous electron and kinetic energy density for
> > the lmbj calculation. What I did is created a new directory and I did
> > the fresh calculation.
> >
> >
> > On Mon, Jul 24, 2023 at 1:22 PM  wrote:
> >>
> >> Could you show how the total energy and distance charge evolve during
> >> the iterations of the lmBJ-SOC calculation (grep for :ENE and DIS in
> >> case.scf)?
> >>
> >> Before using lmBJ-SOC, did you succeed to converge a calculation on your
> >> system using another method, like GGA-PBE or lmBJ without SOC? If yes,
> >> have you tried to start the lmBJ-SOC calculation by using the electron
> >> density (case.clmsum) and kinetic-energy density (case.tausum) obtained
> >> from such a previous calculation that converged?
> >>
> >>
> >> On 22.07.2023 05:47, Burhan Ahmed wrote:
> 

Re: [Wien] calculation with lmbj potential

[Peter Blaha]

Hi,
a) Your vacuum is very large, which can create problems and cost a lot 
of cpu time. 20 Ang are more than enough.
b) As mentioned before: this is a doped SC, so most likely never a gap 
 Is it intended ?


d) Did you relax the slab in PBE ? Just the bulk coordinates are 
probably nonsense.

e) ALWAYS converge PBE first; save_lapw; and use this as start for mBJ.


Am 24.07.2023 um 10:04 schrieb Burhan Ahmed:

The results from the last 20 iterations (for lmbj calculation)

Analysis of parameter:
:ENE :DIS :GAP
in bi2te3lmbj.scf (showing last 20 / 1 lines)

--- ENE ---
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.11941651
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.11305258
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.10571778
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.09359536
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.08522256
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.07976176
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.07273400
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.06627481
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.05653160
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.05078986
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.04165986
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.02824173
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.02513981
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.01487899
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.00743085
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.99627176
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.99306205
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.98317825
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.97811958
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.96939695
--- DIS ---
:DIS  :  CHARGE DISTANCE  (  4.2460241 for atom3 spin 1)  2.0217470
:DIS  :  CHARGE DISTANCE  (  4.2354501 for atom3 spin 1)  2.0167505
:DIS  :  CHARGE DISTANCE  (  4.2249023 for atom3 spin 1)  2.0117662
:DIS  :  CHARGE DISTANCE  (  4.2143828 for atom3 spin 1)  2.0067942
:DIS  :  CHARGE DISTANCE  (  4.2038864 for atom3 spin 1)  2.0018343
:DIS  :  CHARGE DISTANCE  (  4.1934158 for atom3 spin 1)  1.9968867
:DIS  :  CHARGE DISTANCE  (  4.1829711 for atom3 spin 1)  1.9919508
:DIS  :  CHARGE DISTANCE  (  4.1725544 for atom3 spin 1)  1.9870274
:DIS  :  CHARGE DISTANCE  (  4.1621656 for atom3 spin 1)  1.9821164
:DIS  :  CHARGE DISTANCE  (  4.1518018 for atom3 spin 1)  1.9772172
:DIS  :  CHARGE DISTANCE  (  4.1414653 for atom3 spin 1)  1.9723300
:DIS  :  CHARGE DISTANCE  (  4.1311557 for atom3 spin 1)  1.9674546
:DIS  :  CHARGE DISTANCE  (  4.1208686 for atom3 spin 1)  1.9625908
:DIS  :  CHARGE DISTANCE  (  4.1106085 for atom3 spin 1)  1.9577392
:DIS  :  CHARGE DISTANCE  (  4.1003736 for atom3 spin 1)  1.9528994
:DIS  :  CHARGE DISTANCE  (  4.0901658 for atom3 spin 1)  1.9480720
:DIS  :  CHARGE DISTANCE  (  4.0799795 for atom3 spin 1)  1.9432557
:DIS  :  CHARGE DISTANCE  (  4.0698205 for atom3 spin 1)  1.9384514
:DIS  :  CHARGE DISTANCE  (  4.0596869 for atom3 spin 1)  1.9336590
:DIS  :  CHARGE DISTANCE  (  4.0495779 for atom3 spin 1)
1.9288781
#
Yes I have converged the calculation using PBE-GGA and PBE-GGA with
SO. And in both the cases I have found the metallic nature of crystal
(using grep :GAP) and also in band structure the VB crosses the Fermi
level.
##
No, I didn't use the previous electron and kinetic energy density for
the lmbj calculation. What I did is created a new directory and I did
the fresh calculation.


On Mon, Jul 24, 2023 at 1:22 PM  wrote:


Could you show how the total energy and distance charge evolve during
the iterations of the lmBJ-SOC calculation (grep for :ENE and DIS in
case.scf)?

Before using lmBJ-SOC, did you succeed to converge a calculation on your
system using another method, like GGA-PBE or lmBJ without SOC? If yes,
have you tried to start the lmBJ-SOC calculation by using the electron
density (case.clmsum) and kinetic-energy density (case.tausum) obtained
from such a previous calculation that converged?


On 22.07.2023 05:47, Burhan Ahmed wrote:

Dear experts, I am doing an scf calculation taking lmbj potential. My
system is a slab of 6ql with a vacuum of 40 ang along c-axix. Whenever
I try to run the scf calculation including SOC, after 999 iteration
the scf is still not converged. I have analyzed the scf file, it shows
the fluctuating nature. At first the selected Rmt was 2.5 then I
reduces it to 2.34 and again convergence failed after 999 cycle. I
have included -hdlo and -lvns 8 switch in init_lapw as it shows heavy
atom. I am attaching the 

Re: [Wien] calculation with lmbj potential

[Burhan Ahmed]

I did not tried yet (lmbj without SOC).

OK let me try with the same directory where I have converged my pbe-GGA
calculation and let you know.

On Mon, 24 Jul, 2023, 1:47 pm ,  wrote:

> It seems that this calculation was never going to converge. Try to run
> lmBJ-SOC just after GGA-PBE-SOC (and save_lapw) in the same directory.
>
> Is lmBJ without SOC also showing such problems?
>
> On 24.07.2023 10:04, Burhan Ahmed wrote:
> > The results from the last 20 iterations (for lmbj calculation)
> >
> > Analysis of parameter:
> > :ENE :DIS :GAP
> > in bi2te3lmbj.scf (showing last 20 / 1 lines)
> >
> > --- ENE ---
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.11941651
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.11305258
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.10571778
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.09359536
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.08522256
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.07976176
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.07273400
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.06627481
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.05653160
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.05078986
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.04165986
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.02824173
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.02513981
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.01487899
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.00743085
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.99627176
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.99306205
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.98317825
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.97811958
> > :ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.96939695
> > --- DIS ---
> > :DIS  :  CHARGE DISTANCE  (  4.2460241 for atom3 spin 1)
> > 2.0217470
> > :DIS  :  CHARGE DISTANCE  (  4.2354501 for atom3 spin 1)
> > 2.0167505
> > :DIS  :  CHARGE DISTANCE  (  4.2249023 for atom3 spin 1)
> > 2.0117662
> > :DIS  :  CHARGE DISTANCE  (  4.2143828 for atom3 spin 1)
> > 2.0067942
> > :DIS  :  CHARGE DISTANCE  (  4.2038864 for atom3 spin 1)
> > 2.0018343
> > :DIS  :  CHARGE DISTANCE  (  4.1934158 for atom3 spin 1)
> > 1.9968867
> > :DIS  :  CHARGE DISTANCE  (  4.1829711 for atom3 spin 1)
> > 1.9919508
> > :DIS  :  CHARGE DISTANCE  (  4.1725544 for atom3 spin 1)
> > 1.9870274
> > :DIS  :  CHARGE DISTANCE  (  4.1621656 for atom3 spin 1)
> > 1.9821164
> > :DIS  :  CHARGE DISTANCE  (  4.1518018 for atom3 spin 1)
> > 1.9772172
> > :DIS  :  CHARGE DISTANCE  (  4.1414653 for atom3 spin 1)
> > 1.9723300
> > :DIS  :  CHARGE DISTANCE  (  4.1311557 for atom3 spin 1)
> > 1.9674546
> > :DIS  :  CHARGE DISTANCE  (  4.1208686 for atom3 spin 1)
> > 1.9625908
> > :DIS  :  CHARGE DISTANCE  (  4.1106085 for atom3 spin 1)
> > 1.9577392
> > :DIS  :  CHARGE DISTANCE  (  4.1003736 for atom3 spin 1)
> > 1.9528994
> > :DIS  :  CHARGE DISTANCE  (  4.0901658 for atom3 spin 1)
> > 1.9480720
> > :DIS  :  CHARGE DISTANCE  (  4.0799795 for atom3 spin 1)
> > 1.9432557
> > :DIS  :  CHARGE DISTANCE  (  4.0698205 for atom3 spin 1)
> > 1.9384514
> > :DIS  :  CHARGE DISTANCE  (  4.0596869 for atom3 spin 1)
> > 1.9336590
> > :DIS  :  CHARGE DISTANCE  (  4.0495779 for atom3 spin 1)
> > 1.9288781
> > #
> > Yes I have converged the calculation using PBE-GGA and PBE-GGA with
> > SO. And in both the cases I have found the metallic nature of crystal
> > (using grep :GAP) and also in band structure the VB crosses the Fermi
> > level.
> > ##
> > No, I didn't use the previous electron and kinetic energy density for
> > the lmbj calculation. What I did is created a new directory and I did
> > the fresh calculation.
> >
> >
> > On Mon, Jul 24, 2023 at 1:22 PM  wrote:
> >>
> >> Could you show how the total energy and distance charge evolve during
> >> the iterations of the lmBJ-SOC calculation (grep for :ENE and DIS in
> >> case.scf)?
> >>
> >> Before using lmBJ-SOC, did you succeed to converge a calculation on
> >> your
> >> system using another method, like GGA-PBE or lmBJ without SOC? If yes,
> >> have you tried to start the lmBJ-SOC calculation by using the electron
> >> density (case.clmsum) and kinetic-energy density (case.tausum)
> >> obtained
> >> from such a previous calculation that converged?
> >>
> >>
> >> On 22.07.2023 05:47, Burhan Ahmed wrote:
> >> > Dear experts, I am doing an scf calculation taking lmbj potential. My
> >> > system is a slab of 6ql with a vacuum of 40 ang along c-axix. Whenever
> >> > I try to run the scf calculation including SOC, after 

Re: [Wien] calculation with lmbj potential

[fabien . tran]

It seems that this calculation was never going to converge. Try to run 
lmBJ-SOC just after GGA-PBE-SOC (and save_lapw) in the same directory.


Is lmBJ without SOC also showing such problems?

On 24.07.2023 10:04, Burhan Ahmed wrote:

The results from the last 20 iterations (for lmbj calculation)

Analysis of parameter:
:ENE :DIS :GAP
in bi2te3lmbj.scf (showing last 20 / 1 lines)

--- ENE ---
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.11941651
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.11305258
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.10571778
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.09359536
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.08522256
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.07976176
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.07273400
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.06627481
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.05653160
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.05078986
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.04165986
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.02824173
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.02513981
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.01487899
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.00743085
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.99627176
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.99306205
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.98317825
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.97811958
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.96939695
--- DIS ---
:DIS  :  CHARGE DISTANCE  (  4.2460241 for atom3 spin 1)  
2.0217470
:DIS  :  CHARGE DISTANCE  (  4.2354501 for atom3 spin 1)  
2.0167505
:DIS  :  CHARGE DISTANCE  (  4.2249023 for atom3 spin 1)  
2.0117662
:DIS  :  CHARGE DISTANCE  (  4.2143828 for atom3 spin 1)  
2.0067942
:DIS  :  CHARGE DISTANCE  (  4.2038864 for atom3 spin 1)  
2.0018343
:DIS  :  CHARGE DISTANCE  (  4.1934158 for atom3 spin 1)  
1.9968867
:DIS  :  CHARGE DISTANCE  (  4.1829711 for atom3 spin 1)  
1.9919508
:DIS  :  CHARGE DISTANCE  (  4.1725544 for atom3 spin 1)  
1.9870274
:DIS  :  CHARGE DISTANCE  (  4.1621656 for atom3 spin 1)  
1.9821164
:DIS  :  CHARGE DISTANCE  (  4.1518018 for atom3 spin 1)  
1.9772172
:DIS  :  CHARGE DISTANCE  (  4.1414653 for atom3 spin 1)  
1.9723300
:DIS  :  CHARGE DISTANCE  (  4.1311557 for atom3 spin 1)  
1.9674546
:DIS  :  CHARGE DISTANCE  (  4.1208686 for atom3 spin 1)  
1.9625908
:DIS  :  CHARGE DISTANCE  (  4.1106085 for atom3 spin 1)  
1.9577392
:DIS  :  CHARGE DISTANCE  (  4.1003736 for atom3 spin 1)  
1.9528994
:DIS  :  CHARGE DISTANCE  (  4.0901658 for atom3 spin 1)  
1.9480720
:DIS  :  CHARGE DISTANCE  (  4.0799795 for atom3 spin 1)  
1.9432557
:DIS  :  CHARGE DISTANCE  (  4.0698205 for atom3 spin 1)  
1.9384514
:DIS  :  CHARGE DISTANCE  (  4.0596869 for atom3 spin 1)  
1.9336590

:DIS  :  CHARGE DISTANCE  (  4.0495779 for atom3 spin 1)
1.9288781
#
Yes I have converged the calculation using PBE-GGA and PBE-GGA with
SO. And in both the cases I have found the metallic nature of crystal
(using grep :GAP) and also in band structure the VB crosses the Fermi
level.
##
No, I didn't use the previous electron and kinetic energy density for
the lmbj calculation. What I did is created a new directory and I did
the fresh calculation.


On Mon, Jul 24, 2023 at 1:22 PM  wrote:


Could you show how the total energy and distance charge evolve during
the iterations of the lmBJ-SOC calculation (grep for :ENE and DIS in
case.scf)?

Before using lmBJ-SOC, did you succeed to converge a calculation on 
your

system using another method, like GGA-PBE or lmBJ without SOC? If yes,
have you tried to start the lmBJ-SOC calculation by using the electron
density (case.clmsum) and kinetic-energy density (case.tausum) 
obtained

from such a previous calculation that converged?


On 22.07.2023 05:47, Burhan Ahmed wrote:
> Dear experts, I am doing an scf calculation taking lmbj potential. My
> system is a slab of 6ql with a vacuum of 40 ang along c-axix. Whenever
> I try to run the scf calculation including SOC, after 999 iteration
> the scf is still not converged. I have analyzed the scf file, it shows
> the fluctuating nature. At first the selected Rmt was 2.5 then I
> reduces it to 2.34 and again convergence failed after 999 cycle. I
> have included -hdlo and -lvns 8 switch in init_lapw as it shows heavy
> atom. I am attaching the case.struct file. Hoping for any
> suggestion/solution.
>
> Regards
>
> Burhan Ahmed
>
> Research Scholar, AUS
> 

Re: [Wien] calculation with lmbj potential

[Burhan Ahmed]

That is very true. I made the slab using the tutorial available in the wien2k user manual by executing x-supercell program. What I found is that the 6ql supercell consist of an extra Te atom. But whenever I try to remove this atom I got multiplicity error. Sir, what is best possible way of making slab with vacuum or surfaces?? Thanks RegardsBurhan AhmedResearch Scholar, AUS  From: Laurence MarksSent: Saturday, July 22, 2023 6:30 PMTo: A Mailing list for WIEN2k usersSubject: Re: [Wien] calculation with lmbj potential Others will probably give you suggestions about converging mBJ. Some deeper comments. The most common reason that calculations behave badly is user error. Sometimes this is doing the initialization wrong, often it is creating an inappropriate model. Just because one can use x supercell or related codes to create atomic positions does not make them sensible. Your slab has a composition Bi12 Te19, i.e. it is Te rich. You therefore have a reduced, n-type semiconductor with a small gap (about 0.1eV) which will behave badly. If you look at your BVS you will see that atom Te7 is severely underco-ordinated (unstable). I doubt that your slab will ever converge to anything which should be published. Surfaces are not simple, it is unfortunately too easy to set something up and get it wrong. (There are many papers in the literature of dubious merit.) On Fri, Jul 21, 2023 at 10:47 PM Burhan Ahmed <burhan.ah...@aus.ac.in> wrote:Dear experts, I am doing an scf calculation taking lmbj potential. My system is a slab of 6ql with a vacuum of 40 ang along c-axix. Whenever I try to run the scf calculation including SOC, after 999 iteration the scf is still not converged. I have analyzed the scf file, it shows the fluctuating nature. At first the selected Rmt was 2.5 then I reduces it to 2.34 and again convergence failed after 999 cycle. I have included -hdlo and -lvns 8 switch in init_lapw as it shows heavy atom. I am attaching the case.struct file. Hoping for any suggestion/solution.  RegardsBurhan AhmedResearch Scholar, AUS  ___Wien mailing listWien@zeus.theochem.tuwien.ac.athttp://zeus.theochem.tuwien.ac.at/mailman/listinfo/wienSEARCH the MAILING-LIST at:  http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html -- Professor Laurence Marks (Laurie)Department of Materials Science and EngineeringNorthwestern Universitywww.numis.northwestern.edu"Research is to see what everybody else has seen, and to think what nobody else has thought", Albert Szent-Györgyi 
___
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] calculation with lmbj potential

[Burhan Ahmed]

The results from the last 20 iterations (for lmbj calculation)

Analysis of parameter:
:ENE :DIS :GAP
in bi2te3lmbj.scf (showing last 20 / 1 lines)

--- ENE ---
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.11941651
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.11305258
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.10571778
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.09359536
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.08522256
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.07976176
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.07273400
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.06627481
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.05653160
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.05078986
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.04165986
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.02824173
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.02513981
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.01487899
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775750.00743085
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.99627176
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.99306205
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.98317825
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.97811958
:ENE  : *WARNING** TOTAL ENERGY IN Ry =  -775749.96939695
--- DIS ---
:DIS  :  CHARGE DISTANCE  (  4.2460241 for atom3 spin 1)  2.0217470
:DIS  :  CHARGE DISTANCE  (  4.2354501 for atom3 spin 1)  2.0167505
:DIS  :  CHARGE DISTANCE  (  4.2249023 for atom3 spin 1)  2.0117662
:DIS  :  CHARGE DISTANCE  (  4.2143828 for atom3 spin 1)  2.0067942
:DIS  :  CHARGE DISTANCE  (  4.2038864 for atom3 spin 1)  2.0018343
:DIS  :  CHARGE DISTANCE  (  4.1934158 for atom3 spin 1)  1.9968867
:DIS  :  CHARGE DISTANCE  (  4.1829711 for atom3 spin 1)  1.9919508
:DIS  :  CHARGE DISTANCE  (  4.1725544 for atom3 spin 1)  1.9870274
:DIS  :  CHARGE DISTANCE  (  4.1621656 for atom3 spin 1)  1.9821164
:DIS  :  CHARGE DISTANCE  (  4.1518018 for atom3 spin 1)  1.9772172
:DIS  :  CHARGE DISTANCE  (  4.1414653 for atom3 spin 1)  1.9723300
:DIS  :  CHARGE DISTANCE  (  4.1311557 for atom3 spin 1)  1.9674546
:DIS  :  CHARGE DISTANCE  (  4.1208686 for atom3 spin 1)  1.9625908
:DIS  :  CHARGE DISTANCE  (  4.1106085 for atom3 spin 1)  1.9577392
:DIS  :  CHARGE DISTANCE  (  4.1003736 for atom3 spin 1)  1.9528994
:DIS  :  CHARGE DISTANCE  (  4.0901658 for atom3 spin 1)  1.9480720
:DIS  :  CHARGE DISTANCE  (  4.0799795 for atom3 spin 1)  1.9432557
:DIS  :  CHARGE DISTANCE  (  4.0698205 for atom3 spin 1)  1.9384514
:DIS  :  CHARGE DISTANCE  (  4.0596869 for atom3 spin 1)  1.9336590
:DIS  :  CHARGE DISTANCE  (  4.0495779 for atom3 spin 1)
1.9288781
#
Yes I have converged the calculation using PBE-GGA and PBE-GGA with
SO. And in both the cases I have found the metallic nature of crystal
(using grep :GAP) and also in band structure the VB crosses the Fermi
level.
##
No, I didn't use the previous electron and kinetic energy density for
the lmbj calculation. What I did is created a new directory and I did
the fresh calculation.


On Mon, Jul 24, 2023 at 1:22 PM  wrote:
>
> Could you show how the total energy and distance charge evolve during
> the iterations of the lmBJ-SOC calculation (grep for :ENE and DIS in
> case.scf)?
>
> Before using lmBJ-SOC, did you succeed to converge a calculation on your
> system using another method, like GGA-PBE or lmBJ without SOC? If yes,
> have you tried to start the lmBJ-SOC calculation by using the electron
> density (case.clmsum) and kinetic-energy density (case.tausum) obtained
> from such a previous calculation that converged?
>
>
> On 22.07.2023 05:47, Burhan Ahmed wrote:
> > Dear experts, I am doing an scf calculation taking lmbj potential. My
> > system is a slab of 6ql with a vacuum of 40 ang along c-axix. Whenever
> > I try to run the scf calculation including SOC, after 999 iteration
> > the scf is still not converged. I have analyzed the scf file, it shows
> > the fluctuating nature. At first the selected Rmt was 2.5 then I
> > reduces it to 2.34 and again convergence failed after 999 cycle. I
> > have included -hdlo and -lvns 8 switch in init_lapw as it shows heavy
> > atom. I am attaching the case.struct file. Hoping for any
> > suggestion/solution.
> >
> > Regards
> >
> > Burhan Ahmed
> >
> > Research Scholar, AUS
> > ___
> > Wien mailing list
> > Wien@zeus.theochem.tuwien.ac.at
> > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> > SEARCH the MAILING-LIST at:
> > 

Re: [Wien] calculation with lmbj potential

[fabien . tran]

Could you show how the total energy and distance charge evolve during 
the iterations of the lmBJ-SOC calculation (grep for :ENE and DIS in 
case.scf)?


Before using lmBJ-SOC, did you succeed to converge a calculation on your 
system using another method, like GGA-PBE or lmBJ without SOC? If yes, 
have you tried to start the lmBJ-SOC calculation by using the electron 
density (case.clmsum) and kinetic-energy density (case.tausum) obtained 
from such a previous calculation that converged?



On 22.07.2023 05:47, Burhan Ahmed wrote:

Dear experts, I am doing an scf calculation taking lmbj potential. My
system is a slab of 6ql with a vacuum of 40 ang along c-axix. Whenever
I try to run the scf calculation including SOC, after 999 iteration
the scf is still not converged. I have analyzed the scf file, it shows
the fluctuating nature. At first the selected Rmt was 2.5 then I
reduces it to 2.34 and again convergence failed after 999 cycle. I
have included -hdlo and -lvns 8 switch in init_lapw as it shows heavy
atom. I am attaching the case.struct file. Hoping for any
suggestion/solution.

Regards

Burhan Ahmed

Research Scholar, AUS
___
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] calculation with lmbj potential

[Laurence Marks]

Others will probably give you suggestions about converging mBJ. Some deeper
comments.

The most common reason that calculations behave badly is user error.
Sometimes this is doing the initialization wrong, often it is creating an
inappropriate model. Just because one can use x supercell or related codes
to create atomic positions does *not* make them sensible.

Your slab has a composition Bi12 Te19, i.e. it is Te rich. You therefore
have a reduced, n-type semiconductor with a small gap (about 0.1eV) which
will behave badly. If you look at your BVS you will see that atom Te7 is
severely underco-ordinated (unstable).

I doubt that your slab will ever converge to anything which should be
published. Surfaces are not simple, it is unfortunately too easy to set
something up and get it wrong. (There are many papers in the literature of
dubious merit.)

On Fri, Jul 21, 2023 at 10:47 PM Burhan Ahmed 
wrote:

> Dear experts, I am doing an scf calculation taking lmbj potential. My
> system is a slab of 6ql with a vacuum of 40 ang along c-axix. Whenever I
> try to run the scf calculation including SOC, after 999 iteration the scf
> is still not converged. I have analyzed the scf file, it shows the
> fluctuating nature. At first the selected Rmt was 2.5 then I reduces it to
> 2.34 and again convergence failed after 999 cycle. I have included -hdlo
> and -lvns 8 switch in init_lapw as it shows heavy atom. I am attaching the
> case.struct file. Hoping for any suggestion/solution.
>
>
>
> Regards
>
> Burhan Ahmed
>
> *Research Scholar, AUS *
>
>
> ___
> 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
>


-- 
Professor Laurence Marks (Laurie)
Department of Materials Science and Engineering
Northwestern University
www.numis.northwestern.edu
"Research is to see what everybody else has seen, and to think what nobody
else has thought", Albert Szent-Györgyi
___
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