[Wien] Large forces appeared when calculating total energy after structure optimization

[Gavin Abo]

It might also be worth mentioning that you may need to look into the 
details of the structure that you are calculating.

By that, I'm referring to how optimization of atomic coordinates does 
not work on structures were there are no atoms in a free atomic position 
and all atoms in the structure are in fixed atomic positions of the 
spacegroup, like the example given for NaCl, as described in the 
following five links.

[1] Slide 36: 
http://www.wien2k.at/reg_user/textbooks/WIEN2k_lecture-notes_2013/WIEN2k-getting_started.pdf
[2] 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg20232.html
[3] 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg20213.html
[4] 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg02859.html
[5] 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg17074.html

On 2/6/2021 9:40 AM, Peter Blaha wrote:
> All we can say is that you have done somewhere a mistake.
>
> It is important that you not just "describe" what you did, but list 
> the actual commands.
>
> Did you use the same   runsp command during "optimization" ?
>
> Could it be that you restarted the final SCF from "scratch" ???
> In DFT+U you might have ended up in a different state when starting 
> with dstart densities without *dmat* files.
>
> Compare your total energies of best optimization and the final SCF. 
> Which one is lower ?
>
> Am 06.02.2021 um 12:47 schrieb shamik chakrabarti:
>> Dear Wien2k users,
>>                                  I have initially optimized an 
>> orthorhombic crystal system structurally (simultaneous 
>> optimization of a,b,c & coordinates & then optimization of 
>> coordinates again by taking optimized lattice parameters). However, 
>> after taking the optimized structure when I simulate the total energy 
>> with converged SCF by running the command runsp_lapw -dm -orb -fc 1.0 
>> -ec 0.0001 -cc 0.0001, I have found that large forces ~ 40 mRy/a.u.  
>> appeared for few atoms.
>>
>> My queries are (a) why this happens? & (b) what is the remedy? (c) 
>> whether DOS is reliable in this case?
>>
>> *However, the simulated voltage as calculated from the total energy 
>> of the system has been matched with the experiment very well.
>> *
>>
>> Looking forward to hearing from you.
>>
>> with regards,
>> -- 
>> Dr. Shamik Chakrabarti
>> Research Fellow
>> Department of Physics
>> Indian Institute of Technology Patna
>> Bihta-801103
>> Patna
>> Bihar, India