Conflict in parameters of unit cell after CELL_OPTIMIZATION
Mengzi Zhou
smal... at gmail.com
Tue Dec 5 08:54:11 UTC 2017
Hi, Amin
I used CELL_OPT for other minerals and the error compared to experiment's
is always below 1%. I guess it's because of the size of the periodic box.
The length of x, y, z should better be larger than 10A for the accuracy of
calculation in cp2k. I suggest using 3*3*1 supercell for this bulk CELL_OPT
process.
在 2017年11月28日星期二 UTC+8下午5:17:33，Amin写道：
>
> Hi Mengzi,
>
> Thanks for the input. I actually tried with your attached supercell
> coordinates and I got these results: a=b=7.330 and c=10.314 which they have
> 3% and 8% error respectively (in comparison to experimental data). I think
> these amount of error is something needs to be addressed and I'm looking
> for the solution. for the record, here is my input file for my last
> calculation:
>
> &GLOBAL
> PROJECT ana
> RUN_TYPE CELL_OPT
> PRINT_LEVEL LOW
> EXTENDED_FFT_LENGTHS
> &END GLOBAL
>
> &FORCE_EVAL
> METHOD Quickstep
> STRESS_TENSOR ANALYTICAL
> &DFT
> BASIS_SET_FILE_NAME ./BASIS_MOLOPT
> POTENTIAL_FILE_NAME ./GTH_POTENTIALS
> UKS
> WFN_RESTART_FILE_NAME
> MULTIPLICITY 1
> &QS
> METHOD GPW
> &END QS
> &MGRID
> CUTOFF 800
> REL_CUTOFF 60
> &END MGRID
> &SCF
> EPS_SCF 1.0E-6
> MAX_SCF 80
> &OUTER_SCF
> EPS_SCF 1.0E-6
> MAX_SCF 40
> &END
> &OT
> PRECONDITIONER FULL_SINGLE
> MINIMIZER CG
> &END OT
> SCF_GUESS ATOMIC
> &END SCF
> &XC
> &XC_FUNCTIONAL PBE
> &END XC_FUNCTIONAL
> &vdW_POTENTIAL
> DISPERSION_FUNCTIONAL PAIR_POTENTIAL
> &PAIR_POTENTIAL
> TYPE DFTD3(BJ)
> REFERENCE_FUNCTIONAL PBE
> CALCULATE_C9_TERM .FALSE.
> PARAMETER_FILE_NAME ./dftd3.dat
> R_CUTOFF 15.0
> &END PAIR_POTENTIAL
> &END vdW_POTENTIAL
> &END XC
> &PRINT
> &HIRSHFELD
> SELF_CONSISTENT .TRUE.
> &END HIRSHFELD
> &END PRINT
> &POISSON
> PERIODIC XYZ
> POISSON_SOLVER PERIODIC
> &END POISSON
> &END DFT
>
> &SUBSYS
> &CELL
> ABC 7.569 7.569 9.514
> PERIODIC XYZ
> &END CELL
> &COORD
> @INCLUDE 'ana.coord'
> &END COORD
> &KIND O
> BASIS_SET DZVP-MOLOPT-GTH
> POTENTIAL GTH-PBE-q6
> &END KIND
> &KIND Al
> BASIS_SET DZVP-MOLOPT-SR-GTH
> POTENTIAL GTH-PBE-q3
> &END KIND
> &KIND Ti
> BASIS_SET DZVP-MOLOPT-SR-GTH
> POTENTIAL GTH-PBE-q12
> &END KIND
> &END SUBSYS
> &END FORCE_EVAL
>
> &MOTION
> &GEO_OPT
> TYPE MINIMIZATION
> OPTIMIZER BFGS
> &END GEO_OPT
> &CELL_OPT
> KEEP_SYMMETRY TRUE
> KEEP_ANGLES TRUE
> &END
> &END MOTION
>
> Thanks for the consideration.
>
> Amin
>
>
>
> On Monday, November 27, 2017 at 1:58:00 PM UTC+1, Mengzi Zhou wrote:
>>
>> I agree with Matt. The initial structure is not right. You can find the
>> 2*2*1 48 atoms supercell in the attachment, with a=7.56900, b=7.56900,
>> c=9.51430. I use Materials Studio and VMD to generate models and .cif files
>> of minerals can be found on this website
>> http://rruff.geo.arizona.edu/AMS/amcsd.php
>>
>> 在 2017年11月16日星期四 UTC+8下午9:18:17，Amin写道：
>>>
>>> Hi,
>>>
>>> My final intention is to make a slab, but yes I tried with bulk
>>> structure at first. I tried with a 2*2*1 super-cell (containing 48 atoms)
>>> for a bulk Anatase and after cell optimization I got a=7.330, b=7.332 and
>>> c=10.312 which in comparison to experimental data it has 3% error for a(b)
>>> and 8% error for c. I also tried for a unitcell of containing 31 atoms
>>> which I encountered an error saying "Cholesky decompose failed: the matrix
>>> is not positive definite or ill-conditioned".
>>> I have attached both initial structures here.
>>> thanks.
>>>
>>> Amin
>>> On Wednesday, November 15, 2017 at 4:42:10 PM UTC+1, Matt W wrote:
>>>>
>>>> I think your initial structure is strange (wrong). Check everything
>>>> looks correct in an appropriate visualisation package - check all atoms are
>>>> present, maybe some symmetry equivalent ones are missing, for instance.
>>>>
>>>> Are you trying to make a slab, or bulk. Start with bulk.
>>>>
>>>> Matt
>>>>
>>>> On Wednesday, November 15, 2017 at 12:51:24 AM UTC, Amin wrote:
>>>>>
>>>>> I took the initial unit cell from VNL's crystallographic database for
>>>>> Anatase TiO2 and then I made the initial unit cell for Anatase (101) slab.
>>>>> I ran the calculation for Cell Optimization and finally got the results
>>>>> that I mentioned already. Maybe the procedure that I'm employing for making
>>>>> the initial unit cell is not correct?
>>>>>
>>>>>
>>>>>
>>>>> On Tuesday, November 14, 2017 at 11:02:29 PM UTC+1, Matt W wrote:
>>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> I still don't really see anything very wrong with the setup (it is a
>>>>>> bit inefficient, use smaller MAX_SCF in inner loop and FULL_ALL or
>>>>>> FULL_SINGLE_INVERSE preconditioner).
>>>>>>
>>>>>> The initial structure seems very bad - are your coordinates/cell
>>>>>> params correct? Perhaps you converge to a different polymorph or a local
>>>>>> minima?
>>>>>>
>>>>>> Matt
>>>>>>
>>>>>> On Tuesday, November 14, 2017 at 12:14:09 PM UTC, Amin wrote:
>>>>>>>
>>>>>>> Hi Matt,
>>>>>>>
>>>>>>> Thanks for the feedback.
>>>>>>> My results for a,b and c are 11.520*9.789*11.379 but the results in
>>>>>>> the paper are 10.227*11.346*10.540 and yes my SCF converges.
>>>>>>> I have tested with other structures but I couldn't achieve the same
>>>>>>> results which have been done with the same criteria with CP2K. that's why I
>>>>>>> thought there might be something wrong with my input file.
>>>>>>>
>>>>>>> Best,
>>>>>>>
>>>>>>> Amin
>>>>>>>
>>>>>>>
>>>>>>> On Tuesday, November 14, 2017 at 12:45:57 PM UTC+1, Matt W wrote:
>>>>>>>>
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> what do you mean when you say that the results don't agree?
>>>>>>>> Massively wrong, or a small difference?
>>>>>>>>
>>>>>>>> Does your SCF converge? I don't see any very obvious problems with
>>>>>>>> your input at a quick glance.
>>>>>>>>
>>>>>>>> Matt
>>>>>>>>
>>>>>>>> On Tuesday, November 14, 2017 at 11:43:16 AM UTC+1, Amin wrote:
>>>>>>>>>
>>>>>>>>> Dear all,
>>>>>>>>> I'm having problem with my CELL_OPT calculations. I tried to
>>>>>>>>> achieve lattice parameters for Anatase (101) unit cell but my results are
>>>>>>>>> very different than the other results coming from literature. Specially the
>>>>>>>>> results mentioned in the reference paper of this exercise:
>>>>>>>>> https://www.cp2k.org/exercises:2015_pitt:gga
>>>>>>>>> I tried with/without the repetition of the unit cell but the final
>>>>>>>>> lattice parameters was not consistent with those mentioned in literature.
>>>>>>>>> I would be grateful if someone takes a look to my input file and
>>>>>>>>> reminds me what's wrong with my calculations. thanks.
>>>>>>>>>
>>>>>>>>> &GLOBAL
>>>>>>>>> PROJECT cell3
>>>>>>>>> RUN_TYPE CELL_OPT
>>>>>>>>> PRINT_LEVEL LOW
>>>>>>>>> EXTENDED_FFT_LENGTHS
>>>>>>>>> &END GLOBAL
>>>>>>>>>
>>>>>>>>> &FORCE_EVAL
>>>>>>>>> METHOD Quickstep
>>>>>>>>> STRESS_TENSOR ANALYTICAL
>>>>>>>>> &DFT
>>>>>>>>> BASIS_SET_FILE_NAME ./BASIS_MOLOPT
>>>>>>>>> POTENTIAL_FILE_NAME ./GTH_POTENTIALS
>>>>>>>>> UKS
>>>>>>>>> WFN_RESTART_FILE_NAME s-RESTART.wfn
>>>>>>>>> MULTIPLICITY 1
>>>>>>>>> &QS
>>>>>>>>> METHOD GPW
>>>>>>>>> &END QS
>>>>>>>>> &MGRID
>>>>>>>>> CUTOFF 800
>>>>>>>>> REL_CUTOFF 60
>>>>>>>>> &END MGRID
>>>>>>>>> &SCF
>>>>>>>>> EPS_SCF 1.0E-6
>>>>>>>>> MAX_SCF 150
>>>>>>>>> &OUTER_SCF
>>>>>>>>> EPS_SCF 1.0E-6
>>>>>>>>> MAX_SCF 40
>>>>>>>>> &END
>>>>>>>>> &OT
>>>>>>>>> PRECONDITIONER FULL_SINGLE
>>>>>>>>> MINIMIZER CG
>>>>>>>>> &END OT
>>>>>>>>> SCF_GUESS ATOMIC
>>>>>>>>> &END SCF
>>>>>>>>> &XC
>>>>>>>>> &XC_FUNCTIONAL PBE
>>>>>>>>> &END XC_FUNCTIONAL
>>>>>>>>> &vdW_POTENTIAL
>>>>>>>>> DISPERSION_FUNCTIONAL PAIR_POTENTIAL
>>>>>>>>> &PAIR_POTENTIAL
>>>>>>>>> TYPE DFTD3(BJ)
>>>>>>>>> REFERENCE_FUNCTIONAL PBE
>>>>>>>>> CALCULATE_C9_TERM .FALSE.
>>>>>>>>> PARAMETER_FILE_NAME ./dftd3.dat
>>>>>>>>> R_CUTOFF 15.0
>>>>>>>>> &END PAIR_POTENTIAL
>>>>>>>>> &END vdW_POTENTIAL
>>>>>>>>> &END XC
>>>>>>>>> &PRINT
>>>>>>>>> &HIRSHFELD
>>>>>>>>> SELF_CONSISTENT .TRUE.
>>>>>>>>> &END HIRSHFELD
>>>>>>>>> &END PRINT
>>>>>>>>> &POISSON
>>>>>>>>> PERIODIC XYZ
>>>>>>>>> POISSON_SOLVER PERIODIC
>>>>>>>>> &END POISSON
>>>>>>>>> &END DFT
>>>>>>>>>
>>>>>>>>> &SUBSYS
>>>>>>>>> &CELL
>>>>>>>>> ABC 10.2394 3.7845 3.513
>>>>>>>>> PERIODIC XYZ
>>>>>>>>> MULTIPLE_UNIT_CELL 1 3 3
>>>>>>>>> &END CELL
>>>>>>>>> &TOPOLOGY
>>>>>>>>> MULTIPLE_UNIT_CELL 1 3 3
>>>>>>>>> &END TOPOLOGY
>>>>>>>>> &COORD
>>>>>>>>> @INCLUDE 'cell3.coord'
>>>>>>>>> &END COORD
>>>>>>>>> &KIND O
>>>>>>>>> BASIS_SET DZVP-MOLOPT-GTH
>>>>>>>>> POTENTIAL GTH-PBE-q6
>>>>>>>>> &END KIND
>>>>>>>>> &KIND Ti
>>>>>>>>> BASIS_SET DZVP-MOLOPT-SR-GTH
>>>>>>>>> POTENTIAL GTH-PBE-q12
>>>>>>>>> &END KIND
>>>>>>>>> &END SUBSYS
>>>>>>>>> &END FORCE_EVAL
>>>>>>>>>
>>>>>>>>> &MOTION
>>>>>>>>> &GEO_OPT
>>>>>>>>> TYPE MINIMIZATION
>>>>>>>>> OPTIMIZER BFGS
>>>>>>>>> &END GEO_OPT
>>>>>>>>> &CELL_OPT
>>>>>>>>> KEEP_SYMMETRY TRUE
>>>>>>>>> KEEP_ANGLES TRUE
>>>>>>>>> &END
>>>>>>>>> &END MOTION
>>>>>>>>>
>>>>>>>>> cell3.coord:
>>>>>>>>>
>>>>>>>>> O 7.692165e-01 -3.336155e-17 1.170579e+00 0.07512
>>>>>>>>> -0.00000 0.33321
>>>>>>>>> O 5.888893e+00 1.892250e+00 1.170579e+00 0.57512
>>>>>>>>> 0.50000 0.33321
>>>>>>>>> Ti 4.049520e+00 1.892250e+00 1.902225e+00 0.39549
>>>>>>>>> 0.50000 0.54148
>>>>>>>>> Ti 9.169196e+00 -3.976756e-16 1.902225e+00 0.89549
>>>>>>>>> -0.00000 0.54148
>>>>>>>>> O 3.678746e+00 -1.595502e-16 2.049708e+00 0.35928
>>>>>>>>> -0.00000 0.58346
>>>>>>>>> O 8.798422e+00 1.892250e+00 2.049708e+00 0.85928
>>>>>>>>> 0.50000 0.58346
>>>>>>>>> O 2.210147e+00 1.892250e+00 2.633872e+00 0.21585
>>>>>>>>> 0.50000 0.74975
>>>>>>>>> O 7.329824e+00 -3.179005e-16 2.633872e+00 0.71585
>>>>>>>>> -0.00000 0.74975
>>>>>>>>> Ti 1.839373e+00 -7.977512e-17 2.781355e+00 0.17964
>>>>>>>>> -0.00000 0.79173
>>>>>>>>> Ti 6.959049e+00 1.892250e+00 2.781355e+00 0.67964
>>>>>>>>> 0.50000 0.79173
>>>>>>>>> O 0.000000e+00 0.000000e+00 3.513001e+00 0.00000
>>>>>>>>> 0.00000 1.00000
>>>>>>>>> O 5.119676e+00 1.892250e+00 3.513001e+00 0.50000
>>>>>>>>> 0.50000 1.00000
>>>>>>>>>
>>>>>>>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20171205/3c592c3d/attachment.htm>
More information about the CP2K-user
mailing list