[Pw_forum] wrong symmetries for the forces in a supercell (two files enclosed, .in and .out)

[Nicola Marzari]

Dear Manuel,


the only explanation to me is that the code didn't find all the
symmetries that should be there - either because the atomic positions
were not specified with enough accuracy (so the actual unit cell doesn't
have those symmetries) or (maybe?) there is something missing in the
symmetry finder for the rombohedral case. How many operations did the code
find in the ibrav=0 or the ibrav=4 case? A change in lattie constant 
(celldm(1))
wouldn't change the symmetries in your problem.

			nicola


On 30/05/2015 12:40, MKondrin wrote:
> Hi!
>
> I suppose that forces acting on atoms located on positions fixed by
> symmetry might indicate that the lattice constants are not optimal one
> (in fact they depend on pseudopotentials selected). So you should start
> from 'vc-relax' method and then proceed to 'scf' calculations.
>
> Hope this help.
>
> Best regards,
>
> M. V. Kondrin
> (High Pressure Physics Institute, RAS)
>
> Manuel Pérez Jigato wrote:
>>
>> hello
>>
>> I am running an scf calculation for a supercell that is known to have
>> 36 symmetry operations. This supercell (12 atom) contains three
>> primitives of
>> the rhombohedral system CuCrO2 (primitive with 4 atom cell), which is
>> known to have 12 symmetry operations. The supercell is a separate
>> setting (hexagonal setting) of the rhombohedral space group R-3m (ITA
>> number 166) , which is R-3m:H, and, as such, it is a hexagonal system
>>
>> By looking at the Wyckoff positions of my supercell atoms, the site
>> symmetries show that all Cu and all Cr atoms are fixed, ie they do not
>> have any degree of freedom,
>> whilst all the oxygen atoms have their z coordinate not fixed by symmetry
>>
>> The reason for the above explanation is that when I run scf/single
>> point energy calculation for the 12 atom supercell I get forces with
>> wrong symmetries, ie
>>
>>       atom    1 type  1   force =     0.00000000    0.00000000
>>   0.00000000
>>       atom    2 type  1   force =     0.00000000    0.00000000
>> -0.00022475
>>       atom    3 type  1   force =    -0.00000000   -0.00000000
>>   0.00022475
>>       atom    4 type  2   force =    -0.00000000    0.00000000
>>   0.00000000
>>       atom    5 type  2   force =     0.00000000    0.00000000
>>   0.00035717
>>       atom    6 type  2   force =     0.00000000   -0.00000000
>> -0.00035717
>>       atom    7 type  3   force =     0.00000000    0.00000000
>>   0.00241988
>>       atom    8 type  3   force =     0.00000000    0.00000000
>> -0.00241988
>>       atom    9 type  3   force =    -0.00000000    0.00000000
>>   0.00239382
>>       atom   10 type  3   force =     0.00000000    0.00000000
>>   0.00239626
>>       atom   11 type  3   force =     0.00000000    0.00000000
>> -0.00239626
>>       atom   12 type  3   force =     0.00000000   -0.00000000
>> -0.00239382
>>
>> the first 3 lines correspond to Cu, the next 3 lines are Cr, and the
>> last 6 lines are oxygen
>>
>> According to the site symmetries, only the z components of the force
>> on the oxygen atoms (bottom six lines)
>> should be different from zero, but, the output shows that there are z
>> components of the force on Cu and Cr (not all of them) different from
>> zero;
>> as far as I can see, they should be zero
>>
>> This problem appears when I run scf under ibrav 0 and also when I use
>> ibrav 4.
>> After seeing the non-zero forces on Cu and Cr, I have decided to
>> generate the input geometry not by hand, but
>> by means of cif2cell, which gives the ibrav 0 option. Still, the
>> problem of non-zero forces persists.
>> According to the author of cif2cell, the input geometry generated by
>> cif2cell is correct, as well as the CIF file I started from./(See
>> attached file: otra.in)//(See attached file: otra.out)/
>> He suggested that I write to this forum in order to find out about the
>> problem, since he agrees with me on the point that
>> all force components should vanish for all copper and chromium atoms,
>> and that there should be non-zero forces just on all oxygens (z component)
>>
>> will you please help with this? probably some mistake from my side...
>>
>> thanks
>>
>> Manuel
>> PS in order to make sure the FFT grid does not break any symmetries, I
>> run the example with high cutoff, but the same thing happens
>> The k-point set does not break symmetries, since it is Cunningham, ie
>> it contains the gamma-point (odd grid-point number for each of the
>> three lattice directions)
>>
>> Dr Manuel Pérez Jigato, Chargé de Recherche
>> Luxembourg Institute of Science and Technology (LIST)
>> Materials Research and Technology (MRT)
>> 41 rue du Brill
>> L-4422 BELVAUX
>> Grand-Duché de Luxembourg
>> Tel (+352) 47 02 61 - 584
>> Fax (+352) 47 02 64
>> e-mail  manuel.perez at list.lu
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> Pw_forum mailing list
>> Pw_forum at pwscf.org
>> http://pwscf.org/mailman/listinfo/pw_forum
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>

-- 

----------------------------------------------------------------------
Prof Nicola Marzari, Chair of Theory and Simulation of Materials, EPFL
Director, National Centre for Competence in Research NCCR MARVEL, EPFL