[Pw_forum] Fw: SV: SV: SV: cif2cell enquiry TO MKONDRIN ( one file enclosed )
Nicola Marzari
nicola.marzari at epfl.ch
Sun May 31 11:50:01 CEST 2015
Dear Manuel,
Torbjörn is saying exactly the same thing I said, re fractional
translations etc...
In my opiinion, f you want to get exactly zero forces on the atoms that
should have zero forces, monitor them as a function of k-point sampling.
Ss you increase the k-points, the ground state charge density should
converge to the symmetric one, even if it has not been symmetrized by
hand by the code.
Also note that 5 5 1 sampling seems to me very inhomogeneous in the
three different directions - why don't you try 2 2 2, 3 3 3, 4 4 4, 8 8
8, ...
and see how it goes (and maybe report?).
Thanks a lot,
nicola
On 30/05/2015 19:03, Manuel Pérez Jigato wrote:
> hi mkondrin
>
> the email below, from the author of cif2cell, explains the reasons for
> having
> zero forces for all copper and chromium atoms on the basis of the CIF
> file site
> symmetries; please find enclosed the cif file
>
> regards
>
> Manuel
> PS the expected forces, qualitatively speaking are shown below (first 6
> lines are oxygen, next three lines are Cr, last 3 lines are copper
> atoms), from a low cutoff, small number of k-points abinit calculation
> (in the espresso calculation some copper and chromium atoms are fixed,
> other copper and chromium atoms move, which means some of the
> copper-copper, chromium-copper and chromium-chromium distances vary, and
> theyshould all be fixed)
>
> cartesian forces (eV/Angstrom) at end:
> 1 0.00000000000000 0.00000000000000 33.52744537217701
> 2 0.00000000000000 0.00000000000000 33.52744537217701
> 3 0.00000000000000 0.00000000000000 33.52744537217700
> 4 0.00000000000000 0.00000000000000 -33.52744537217701
> 5 0.00000000000000 0.00000000000000 -33.52744537217701
> 6 0.00000000000000 0.00000000000000 -33.52744537217700
> 7 0.00000000000000 0.00000000000000 0.00000000000000
> 8 0.00000000000000 0.00000000000000 0.00000000000000
> 9 0.00000000000000 0.00000000000000 0.00000000000000
> 10 0.00000000000000 0.00000000000000 0.00000000000000
> 11 0.00000000000000 0.00000000000000 0.00000000000000
> 12 0.00000000000000 0.00000000000000 0.00000000000000
>
> 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
>
> -----Forwarded by Manuel Pérez Jigato/LIST on 05/30/2015 06:35PM -----
> To: Manuel Pérez Jigato <manuel.perez at list.lu>
> From: Björkman Torbjörn <torbjorn.bjorkman at aalto.fi>
> Date: 05/29/2015 02:04PM
> Subject: SV: SV: SV: cif2cell enquiry
>
> Well, whether or not it is explicitly used for symmetrization, the
> symmetry is of course still there. So if by symmetry the forces should
> be zero, then they should be zero. In your case I can see from the CIF
> that the forces should indeed be zero, since they are 3a and 3b Wyckoff
> positions, which have no free parameters. If they are not, then there is
> some form of numerical problem, how bad it is of course depends on how
> large the forces are. The manual mentioned that the fourier mesh might
> not having the symmetry of the lattice in all cases, it may be that it
> is possible to get that in order by selecting them as multiples of 3 for
> a trigonal lattice. That's all I can say, I don't know anything specific
> about QE.
>
> Cheers,
> Torbjörn
>
> ---
> Torbjörn Björkman, PhD
> COMP, Aalto University School of Science
> Espoo, Finland
>
> ------------------------------------------------------------------------
> *Från:* Manuel Pérez Jigato [manuel.perez at list.lu]
> *Skickat:* den 29 maj 2015 14:42
> *Till:* Björkman Torbjörn
> *Ämne:* Re: SV: SV: cif2cell enquiry
>
> thanks a lot Bjorkman, it is very kind of you
> my only worry would be the possible implications for the forces
> I will ask to the pwscf mailing list
> regards
> Manuel
>
> Inactive hide details for Björkman Torbjörn ---29/05/2015
> 12:08:34---Dear Manuel, I can find nothing that looks wrong in the
> ciBjörkman Torbjörn ---29/05/2015 12:08:34---Dear Manuel, I can find
> nothing that looks wrong in the cif2cell output, but a quick look in the "Tr
>
> From: Björkman Torbjörn <torbjorn.bjorkman at aalto.fi>
> To: Manuel Pérez Jigato <manuel.perez at list.lu>,
> Date: 29/05/2015 12:08
> Subject: SV: SV: cif2cell enquiry
>
> ------------------------------------------------------------------------
>
>
>
> Dear Manuel,
>
> I can find nothing that looks wrong in the cif2cell output, but a quick
> look in the "Troubleshooting" section of the QE manual seems to provide
> the most likely answer:
>
> pw.x does not find all the symmetries you expected
>
>
> ...
>
> * a fractional translation, without rotation, is a symmetry operation
> of the system. This means that the cell is actually a supercell. In
> this case, all symmetry operations containing fractional
> translations are disabled. The reason is that in this rather exotic
> case there is no simple way to select those symmetry operations
> forming a true group, in the mathematical sense of the term.
>
>
> What they say is that QE in cases like these will just say "this is a
> supercell" and throw the fractional translations away, which gives you
> 12 symmetries instead of 36, just as it should. I have no idea if you
> can bypass this behaviour in QE though, better ask on their mailing list.
>
> Hope this helps,
> Torbjörn
>
> ---
> Torbjörn Björkman, PhD
> COMP, Aalto University School of Science
> Espoo, Finland
>
> ------------------------------------------------------------------------
>
> *Från:* Manuel Pérez Jigato [manuel.perez at list.lu]*
> Skickat:* den 29 maj 2015 12:37*
> Till:* Björkman Torbjörn*
> Ämne:* Re: SV: cif2cell enquiry
>
> Dear Bjorkman
>
> thanks for your prompt reply
> please find enclosed the files/(See attached file: esta.cif)(See
> attached file: QE_NOREDUCE_esta.in)(See attached file: second.in)(See
> attached file: second.out)(See attached file: output_noreduce.txt)/
>
> Manuel
> PS the different between the two .in files is that one of them is the
> direct output from cif2cell, whilst the other is produced by
> adding a few options by hand, before running quantum-espresso (whose
> output you can see as well)
>
>
> _______________________________________________
> 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
More information about the users
mailing list