[Wien] Volume optimization and Eloss function.

[Gavin Abo]

Since I haven't ran many hexagonal optimizations, I probably cannot 
provide as good of advice as some of the others in the list likely could.

My first piece of advice would be to mention which particular package 
you used for your optimization.  optimize, optimize_abc, or 2DRoptimize?

 From the looks of it, it is likely the 2DRoptimize.  In other words, it 
could be easier when the reader doesn't have to guess or infer at what 
was used.

Changing the percentage for altering the structure until you find the 
minimum energy point as you have stated should be the correct procedure.

Looking at the energy vs volume I believe typically worked better for 
cubic structures (e.g. volume of a^3, where a=b=c).

So, you might try looking at energy vs c/a to see if that works better 
or not for finding the minimum energy point for your calculation case.

If optimize was used and option (5) VARY A and C (2D-case) (tetragonal 
or hexagonal lattice) was selected, I believe it will ask for a percent 
change of the c/a ratio.  If I recall correctly, w2web has an energy vs 
c/a plot selection that you could try using when option (5) is selected 
for the optimize program.  I think there is also a parabolfit_lapw 
script that can help analyze the different degrees of freedom for a 
hexagonal structure (2 lattice parameters: a and c).  Based on the 
WIEN2k usersguide, it looks like a and c when using the optimize_abc 
package can be controlled by the -d X argument with X needing to be the 
replaced with the amount of change as a number value (in percent).

It looks like the hexagonal example in the 2Doptimize (pdf) at [1], if 
that is what you were following, takes inputs for both a volume change 
and c/a change as I see in the document:

*******Start for volume changes*******

Do you want to use default parameters  ?(y/n)

(-10%,-5%,0,5%,10%)

y

******* Start for c/a changes*******

(-4%, -2%,0,2%,4%)

y

In the above, the default is -10 to 10% in 5% increments.

 From your description, it looks like you changed the volume from -30% 
to 20%.  However, it doesn't look like you mention which increments you 
took between them. Also, is it safe or dangerous to assume that you took 
the default for the c/a changes?

Let's say a physical atom has a radius, /r/.

If you view the struct file for the structure in XCrySDen, if you see 
/r/, surface of the ball, of one atom (e.g., Hf) touch or overlap the 
/r/ of another atom (e.g., O), then you have a bad struct file.  You 
would have to fix the lattice constants or lattice angles.  For example, 
that might happen if the lattice constants are entered in the struct 
file in the wrong units [2,3].

If you saw the recent post at [4], you might have read how the 
muffin-tin radius (RMT) > /r/ has radial functions less the RMT and 
plane waves greater than RMT.

Perhaps you have also seen Figure 1 at [5] or Figure 1 in [6] 
illustrating the muffin-tin and interstitial regions.

The 'NN' - overlapping spheres error will happen when the RMT of two 
atoms touch or overlap each other.

If you have lattice parameters that you are not needing to optimize [7], 
it is likely straightforward setting 0% in setrmt during initialization 
of the calculation.

However, when doing a lattice parameter optimization you could adjust 
the percentage to reduce the RMT [8] for trying to avoid later on 
overlapping sphere errors.  You might recall when you went through the 
TiC tutorial in the WIEN2k 23.1 usersguide on page 17 [9] it used 3% for 
that example.


[1] http://susi.theochem.tuwien.ac.at/reg_user/textbooks/
[2] 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg06323.html
[3] 
http://zeus.theochem.tuwien.ac.at/pipermail/wien/2010-January/012330.html
[4] 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg22739.html
[5] http://susi.theochem.tuwien.ac.at/lapw/index.html
[6] https://arxiv.org/abs/2004.10728v1
[7] 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg17074.html
[8] http://zeus.theochem.tuwien.ac.at/pipermail/wien/2018-May/028064.html
[9] http://www.wien2k.at/reg_user/textbooks/usersguide.pdf


Hopefully that helps and best regards,
Gavin
WIEN2k user

On 8/22/2023 4:04 PM, Pranjal Nandi wrote:
>
> Dear Sir,
>
> Thank you for the tip. I have used it and am carrying on the 
> calculations. I need further guidance.
>
> As far as my understanding goes, I should keep changing the percentage 
> till the time I don’t see a minimum energy point (a curve whose either 
> side should be higher, but till now I am only getting a straight line) 
> in the plot of energy vs volume. For the first few calculations, I 
> changed the volume from -20 to 20% and found that -20 is has the 
> minimum energy. Then when I change it from -30 to -10 (to see -30 has 
> lower energy or not), I get the error grep: lapw2*./error/: /No such 
> file or directory exists (probably ghost bands are occurring at -30)./
>
> //
>
> /So next what I think I should do is redo the initialisation and run 
> scf with the minimum energy struct file obtained in the earlier 
> optimisation and don’t change the structure suggested by sgroup and 
> again repeat the optimisation on this struct file so that I can check 
> if loweing the volume more than -20 can minimise the energy (I will 
> start with -10 to 0 as the existing structure is already -20). /
>
> //
>
> /However, now I am getting the error 'NN' - overlapping spheres, which 
> I will try to fix it (if free you may comment on this too but I guess 
> I may find a thread addressed to this error). /
>
> //
>
> /*But my question is that am I going in the right direction with the 
> optimisation procedure?*/**
>
> Thank you for your help once again.
>
> With warm regards,
>
> Pranjal
>
> *From:*Wien <wien-bounces at zeus.theochem.tuwien.ac.at> *On Behalf Of 
> *Gavin Abo
> *Sent:* Friday, August 18, 2023 5:06 AM
> *To:* wien at zeus.theochem.tuwien.ac.at
> *Subject:* Re: [Wien] Volume optimization and Eloss function.
>
> If the structure is still hexagonal after the substitution, the 
> starting point could be section "5.3 Structure optimization" on 
> lattice parameter optimization starting on page 87 in the WIEN2k 23.1 
> (or 23.2) usersguide [1].
>
> There you should see that you may select between two different packages.
>
> If using *optimize*, there is:
>
> [5] VARY A and C (2D-case) (tetragonal or hexagonal lattice)
>
> Or if using *optimize_abc_lapw*, there is:
>
> optimize_abc -t 2
>
> After that, if the structure contains any free positions for atomic 
> position optimization, then there is section "5.3.2 Minimization of 
> internal parameters" on page 91 of the usersguide [1].
>
> Yes, case.eloss is for the loss function [2].  On slide 9 in [3], it 
> looks to be showing that case.epsilon contains the real and imaginary 
> parts for the complex dielectric tensor and case.sigmak contains the 
> real and imaginary parts for the optical conductivity.
>
> [1] 
> http://www.wien2k.at/reg_user/textbooks/usersguide.pdfhttp://www.wien2k.at/reg_user/textbooks/usersguide.pdf 
> <http://www.wien2k.at/reg_user/textbooks/usersguide.pdfhttp:/www.wien2k.at/reg_user/textbooks/usersguide.pdf>
> [2] Slide 22: 
> http://www.wien2k.at/events/ws2006/Optics_Vienna_April_2006.pdf
> [3] 
> http://www.wien2k.at/reg_user/textbooks/WIEN2k_lecture-notes_2013/optic_handout.pdf
>
> On 8/17/2023 11:23 AM, Pranjal Nandi wrote:
>
>     Dear All,
>
>     I have an issue with 2 different parts (not related to each
>     other). They are as follows.
>
>      1. I am having a hard time in getting a hcp structure of HfO0.7
>         (or Hf605 which is close to 0.7). Therefore, what I have done
>         is that I have downloaded the hcp structure of HfPo and
>         substituted the Po with Oxygen.
>
>         Now, I want to do the structural and volumetric relaxation so
>         that I have the stable relaxed structure of HfO0.7. I did read
>         the guidebook but I need guidance on what should be the
>         correct order of optimisation (I am lost as I can’t understand
>         from which optimisation I should start).
>
>
>      2. In the eloss programme, which part of the function does the
>         eloss, epsilon and sigmak correspond to ? I guess eloss =
>         energy loss function, epsilon = real part of the dielectric
>         function , sigmaK = imaginary part of the dielectric function.
>         Am I right?
>
>     Looking forward to your kind guidance.
>
>     Thank you.
>
>     With warm regards,
>
>     Pranjal
>
>
>
>     Aquest missatge, i els fitxers adjunts que hi pugui haver, pot
>     contenir informació confidencial o protegida legalment i s’adreça
>     exclusivament a la persona o entitat destinatària. Si no consteu
>     com a destinatari final o no teniu l’encàrrec de rebre’l, no esteu
>     autoritzat a llegir-lo, retenir-lo, modificar-lo, distribuir-lo,
>     copiar-lo ni a revelar-ne el contingut. Si l’heu rebut per error,
>     informeu-ne el remitent i elimineu del sistema tant el missatge
>     com els fitxers adjunts que hi pugui haver.
>
>     Este mensaje, y los ficheros adjuntos que pueda incluir, puede
>     contener información confidencial o legalmente protegida y está
>     exclusivamente dirigido a la persona o entidad destinataria. Si
>     usted no consta como destinatario final ni es la persona encargada
>     de recibirlo, no está autorizado a leerlo, retenerlo, modificarlo,
>     distribuirlo o copiarlo, ni a revelar su contenido. Si lo ha
>     recibido por error, informe de ello al remitente y elimine del
>     sistema tanto el mensaje como los ficheros adjuntos que pueda
>     contener.
>
>     This email message and any attachments it carries may contain
>     confidential or legally protected material and are intended solely
>     for the individual or organization to whom they are addressed. If
>     you are not the intended recipient of this message or the person
>     responsible for processing it, then you are not authorized to
>     read, save, modify, send, copy or disclose any part of it. If you
>     have received the message by mistake, please inform the sender of
>     this and eliminate the message and any attachments it carries from
>     your account.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20230823/b3a4273c/attachment.htm>