[CP2K-user] [CP2K:18923] Re: Incorrect optimized geometry using CELL_OPT
Diego López
diegolopal at gmail.com
Tue Jun 13 12:31:05 UTC 2023
Hi Fabian
I tried using smearing and now the magnetic moments and the PDOS look
really nice! I will try to optimize the system with this set up. Thanks a
lot for your comments.
Best,
Diego
El martes, 13 de junio de 2023 a las 12:32:30 UTC+2, fabia... at gmail.com
escribió:
> Hi Diego,
>
> In QE you have (almost certainly) used semearing too. It can be used with
> semiconductors just as well as with metals. Between each geometry
> optimization step there is an energy minimization, did these all converge?
>
> Cheers
> On Tuesday, 13 June 2023 at 12:24:45 UTC+2 Diego López wrote:
>
>> Hi Fabian,
>>
>> Thanks for your answer. My system is a semiconductor so I think that
>> smearing is not a good choice for my calculations. Using QE I obtained a
>> good approximation of the magnetic moments of the system and also a
>> reasonable geometry optimization. What do you mean with convergence of
>> individual scf steps?
>>
>> Best,
>> Diego
>>
>> El martes, 13 de junio de 2023 a las 12:03:06 UTC+2, fabia... at gmail.com
>> escribió:
>>
>>> Hi Diego,
>>>
>>> I have not used RELAX_MULTIPLICITY myself, you can give it a try. You
>>> could also use DIAGONALIZATION together with SMEAR then the spins are not
>>> fixed by default. What spin population did you obtain with vasp and QE?
>>>
>>> Have you also checked that all the individual scf between the geometry
>>> optimization steps are converged?
>>>
>>> Cheers,
>>> Fabian
>>> On Tuesday, 13 June 2023 at 10:17:28 UTC+2 Diego López wrote:
>>>
>>>> Dear Fabian,
>>>>
>>>> I am using this specific multiplicity because experimental data
>>>> suggests that this is the magnetic ground state, but I am not obtaining the
>>>> proper magnetic moments on the Fe atoms. Should I try using
>>>> RELAX_MULTIPLICITY?
>>>>
>>>> Regards,
>>>> Diego
>>>>
>>>> El sábado, 10 de junio de 2023 a las 19:26:46 UTC+2, Fabian Ducry
>>>> escribió:
>>>>
>>>>> Hi Diego,
>>>>>
>>>>> are you sure that MULTIPLICITY 301 is correct?
>>>>>
>>>>> Cheers,
>>>>> Fabian
>>>>>
>>>>>
>>>>> On 10 June 2023 16:26:05 GMT+01:00, Lobna Saeed <lobna... at gmail.com>
>>>>> wrote:
>>>>>
>>>>>> And also please do energy convergence tests. Dont try to increase
>>>>>> your pw cutoff values randomly.
>>>>>>
>>>>>> On Sat 10. Jun 2023 at 17:25, Lobna Saeed <lobna... at gmail.com> wrote:
>>>>>>
>>>>>>> The minimizer DIIS is not reliable. This OT method is not always
>>>>>>> converging with me and when converges doesnt give accurate enough numbers.
>>>>>>> Therefore, please try to switch to diagonalization.
>>>>>>>
>>>>>>> On Fri 9. Jun 2023 at 16:44, Diego López <diego... at gmail.com> wrote:
>>>>>>>
>>>>>>>> Hi Marcella,
>>>>>>>>
>>>>>>>> Thanks for your answer. I tried with PW cutoff of 1000 and with
>>>>>>>> TZVP basis set but the result is the same
>>>>>>>>
>>>>>>>> Regards
>>>>>>>> Diego
>>>>>>>> El viernes, 9 de junio de 2023 a las 16:38:41 UTC+2, Marcella
>>>>>>>> Iannuzzi escribió:
>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Hi,
>>>>>>>>>
>>>>>>>>> I would start by using better basis sets (e.g., DZVP, TZVP) and
>>>>>>>>> increasing the PW cutoff
>>>>>>>>>
>>>>>>>>> Regards
>>>>>>>>> Marcella
>>>>>>>>>
>>>>>>>>> On Friday, June 9, 2023 at 1:00:42 PM UTC+2 Diego López wrote:
>>>>>>>>>
>>>>>>>>>> Dear CP2K users,
>>>>>>>>>>
>>>>>>>>>> I am kind of new using cp2k but I have been using the software fo
>>>>>>>>>> a couple of months. I am trying to optimize both geometry and cell
>>>>>>>>>> parameters of different semiconductor material but the obtained geometries
>>>>>>>>>> are distorted compared with experimental data or relaxed geometries
>>>>>>>>>> obtained using Quantum Espresso or VASP.
>>>>>>>>>>
>>>>>>>>>> My optimized geometries have a lack of symmetry compared with
>>>>>>>>>> experimental data, i.e. the atoms that should be aligned in an especific
>>>>>>>>>> axis tend to alternate the positions in the crystal, so my results are no
>>>>>>>>>> compatible with the expected ones.
>>>>>>>>>>
>>>>>>>>>> Which parameters should I change in my calculations? I tried
>>>>>>>>>> lowering the EPS_SCF to increase accuracy, increasing the supercell and
>>>>>>>>>> using larger basis sets but the result does not change. I attatch an
>>>>>>>>>> example of input that I used for these calculations.
>>>>>>>>>>
>>>>>>>>>> &GLOBAL
>>>>>>>>>> PROJECT
>>>>>>>>>> RUN_TYPE CELL_OPT
>>>>>>>>>> PRINT_LEVEL MEDIUM
>>>>>>>>>> &END GLOBAL
>>>>>>>>>>
>>>>>>>>>> &FORCE_EVAL
>>>>>>>>>> METHOD Quickstep ! Electronic structure method
>>>>>>>>>> (DFT,...)
>>>>>>>>>> STRESS_TENSOR ANALYTICAL
>>>>>>>>>> &DFT
>>>>>>>>>> BASIS_SET_FILE_NAME BASIS_MOLOPT
>>>>>>>>>> POTENTIAL_FILE_NAME GTH_POTENTIALS
>>>>>>>>>> MULTIPLICITY 301
>>>>>>>>>> UKS
>>>>>>>>>> &POISSON ! Solver requested for non
>>>>>>>>>> periodic calculations
>>>>>>>>>> PERIODIC XYZ
>>>>>>>>>> &END POISSON
>>>>>>>>>> &XC
>>>>>>>>>> &XC_FUNCTIONAL PBE
>>>>>>>>>> &PBE
>>>>>>>>>> PARAMETRIZATION PBESOL
>>>>>>>>>> &END
>>>>>>>>>> &END XC_FUNCTIONAL
>>>>>>>>>> &END XC
>>>>>>>>>> &SCF
>>>>>>>>>> SCF_GUESS ATOMIC
>>>>>>>>>> MAX_SCF 40
>>>>>>>>>> EPS_SCF 3E-06
>>>>>>>>>> &OT
>>>>>>>>>> PRECONDITIONER FULL_KINETIC
>>>>>>>>>> MINIMIZER CG
>>>>>>>>>> &END OT
>>>>>>>>>> &OUTER_SCF
>>>>>>>>>> MAX_SCF 320
>>>>>>>>>> EPS_SCF 3E-06
>>>>>>>>>> &END
>>>>>>>>>> &END SCF
>>>>>>>>>> &MGRID
>>>>>>>>>> NGRIDS 4
>>>>>>>>>> CUTOFF 550
>>>>>>>>>> REL_CUTOFF 90
>>>>>>>>>> &END MGRID
>>>>>>>>>> &QS
>>>>>>>>>> METHOD GPW
>>>>>>>>>> EXTRAPOLATION ASPC
>>>>>>>>>> EPS_DEFAULT 1E-014
>>>>>>>>>> &END
>>>>>>>>>> &END DFT
>>>>>>>>>> &SUBSYS
>>>>>>>>>> &CELL
>>>>>>>>>> ABC 3.601367151 4.818258107 25.0
>>>>>>>>>> ALPHA_BETA_GAMMA 90 90 90
>>>>>>>>>> PERIODIC XYZ
>>>>>>>>>> SYMMETRY ORTHORHOMBIC
>>>>>>>>>> MULTIPLE_UNIT_CELL 5 5 1
>>>>>>>>>> &END CELL
>>>>>>>>>> &TOPOLOGY ! Section used to center the
>>>>>>>>>> atomic coordinates in the given box. Useful for big molecules
>>>>>>>>>> COORD_FILE_FORMAT xyz
>>>>>>>>>> COORD_FILE_NAME ./crsbr.xyz
>>>>>>>>>> MULTIPLE_UNIT_CELL 5 5 1
>>>>>>>>>> &END
>>>>>>>>>> &KIND Cr
>>>>>>>>>> ELEMENT Cr
>>>>>>>>>> BASIS_SET SZV-MOLOPT-SR-GTH
>>>>>>>>>> POTENTIAL GTH-PBE-q14
>>>>>>>>>> MAGNETIZATION 3.0
>>>>>>>>>> &END KIND
>>>>>>>>>> &KIND S
>>>>>>>>>> ELEMENT S
>>>>>>>>>> BASIS_SET SZV-MOLOPT-GTH
>>>>>>>>>> POTENTIAL GTH-PBE-q6
>>>>>>>>>> &END KIND
>>>>>>>>>> &KIND Br
>>>>>>>>>> ELEMENT Br
>>>>>>>>>> BASIS_SET SZV-MOLOPT-SR-GTH
>>>>>>>>>> POTENTIAL GTH-PBE-q7
>>>>>>>>>> &END KIND
>>>>>>>>>> &END SUBSYS
>>>>>>>>>> &END FORCE_EVAL
>>>>>>>>>> &MOTION
>>>>>>>>>> &CELL_OPT
>>>>>>>>>> TYPE DIRECT_CELL_OPT
>>>>>>>>>> MAX_ITER 200
>>>>>>>>>> OPTIMIZER LBFGS
>>>>>>>>>> KEEP_SYMMETRY
>>>>>>>>>> &END CELL_OPT
>>>>>>>>>> &END MOTION
>>>>>>>>>>
>>>>>>>>>> Thanks in advanced
>>>>>>>>>>
>>>>>>>>> --
>>>>>>>> You received this message because you are subscribed to the Google
>>>>>>>> Groups "cp2k" group.
>>>>>>>> To unsubscribe from this group and stop receiving emails from it,
>>>>>>>> send an email to cp2k+uns... at googlegroups.com.
>>>>>>>> To view this discussion on the web visit
>>>>>>>> https://groups.google.com/d/msgid/cp2k/fb6271ad-4513-41a8-a08c-f20b5c23e849n%40googlegroups.com
>>>>>>>> <https://groups.google.com/d/msgid/cp2k/fb6271ad-4513-41a8-a08c-f20b5c23e849n%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>>>>>> .
>>>>>>>>
>>>>>>>
--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/58d65c37-8011-441a-b2c6-a36636da7462n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20230613/9dd01283/attachment.htm>
More information about the CP2K-user
mailing list