[QE-users] Problem with PBE0 calculation
Giuseppe Mattioli
giuseppe.mattioli at ism.cnr.it
Thu Jan 12 11:06:48 CET 2023
Dear M. V. Kondrin
The "scissors operator" strategy can be appropriate in the case of
stoichiometric materials, but it may be not when you investigate
defects or dopants in a host matrix, where the band structure (and the
DOS) can depend on the localization of electronic charge on defects
that can be seen as "deep" or "shallow" by different GGA or EXX
functionals. The best method to use depend on Rameswar's purposes.
Best
Giuseppe
Quoting mkondrin <mkondrin at hppi.troitsk.ru>:
> I also performed calculations using HSE06 functional for obtaining
> electron DOS (using projwfc.x). It is compatible with hybrid
> functionals and usually provide correct values of electron bandgaps.
> After that I just used "scissors" operation on electron band
> structure obtained using standard PBE to match bandgap obtained
> previously.
>
> Approximately in the same words I described this operation in the
> recent paper which is under review now.
>
> Hope this helps.
>
> Sincerely yours,
> M. V. Kondrin
>
>
>
> On 01/12/2023 11:03 AM, Giuseppe Mattioli wrote:
>>
>> AFAIK, nscf calculations are not compatible with EXX functionals.
>> There are a few workarounds, like manually adding the band path as
>> "fake" k-points with weight=0 to the regular grid obtained by the
>> automatic distribution used in the scf calculation.
>> I generally perform EXX calculations of isolated or very large
>> systems @Gamma. More experienced users may add more accurate
>> information.
>> HTH
>> Giuseppe
>>
>> Quoting Rameswar Bhattacharjee <rb1820 at georgetown.edu>:
>>
>>> Hi Giuseppe,
>>> Thank you very much for your detailed reply. I followed the instructions
>>> and was able to complete the SCF calculation. Also, I did an NSCF
>>> calculation using the same combination you suggested as a pre-step to
>>> perform the band calculation. My ultimate aim is to get the band gap using
>>> a hybrid functional such as PBE0.
>>>
>>> However, I am not able to run the band calculation with this combination
>>> got an error like below:
>>>
>>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Error in routine exx_grid_init
>>> (1):
>>>
>>> wrong EXX q grid
>>>
>>>
>>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% It seems the k-points grid I am using is not suitable for EXX
>>> calculation
>>> although I used the same grid for PBE and it was perfect. I am sure someone
>>> in the forum had performed band structure using hybrid functional and would
>>> be able to help me resolve this. Any help would be highly appreciated.
>>>
>>> Here is my input for band calculation using PBE0
>>>
>>> ------------------------------------------
>>>
>>> &control
>>>
>>> restart_mode='from_scratch'
>>>
>>> calculation='bands'
>>>
>>> prefix='Ethyne-Anthracene',
>>>
>>> pseudo_dir = '/home/rb1820/QE-PP/NC',
>>>
>>> outdir='./scratch'
>>>
>>> /
>>>
>>> &system
>>>
>>> ibrav = 0,
>>>
>>> tot_charge=0,
>>>
>>> nat=36,
>>>
>>> ntyp=2,
>>>
>>> ecutwfc = 90,
>>>
>>> ecutfock= 150,
>>>
>>> input_dft='pbe0'
>>>
>>> vdw_corr='grimme-d2',
>>>
>>> nbnd = 60
>>>
>>> /
>>>
>>> &electrons
>>>
>>> electron_maxstep = 1000
>>>
>>> mixing_beta=0.7
>>>
>>> diagonalization='david'
>>>
>>> diago_david_ndim = 2
>>>
>>> conv_thr=1.0d-8
>>>
>>> /
>>>
>>> &ions
>>>
>>> /
>>>
>>> &CELL
>>>
>>> cell_dynamics='bfgs',
>>>
>>> press = 0.0
>>>
>>> cell_dofree = 'x',
>>>
>>> /
>>>
>>> ATOMIC_SPECIES
>>>
>>> H 1.008 H_ONCV_PBE-1.2.upf
>>>
>>> C 12.0107 C_ONCV_PBE-1.2.upf
>>>
>>>
>>> ATOMIC_POSITIONS {angstrom}
>>>
>>> K_POINTS { crystal }
>>>
>>> 11
>>>
>>> -0.5 0.0 0.0 1.0
>>>
>>> -0.4 0.0 0.0 1.0
>>>
>>> -0.3 0.0 0.0 1.0
>>>
>>> -0.2 0.0 0.0 1.0
>>>
>>> -0.1 0.0 0.0 1.0
>>>
>>> 0.0 0.0 0.0 1.0
>>>
>>> 0.1 0.0 0.0 1.0
>>>
>>> 0.2 0.0 0.0 1.0
>>>
>>> 0.3 0.0 0.0 1.0
>>>
>>> 0.4 0.0 0.0 1.0
>>>
>>> 0.5 0.0 0.0 1.0
>>>
>>>
>>> CELL_PARAMETERS {angstrom}
>>>
>>> 6.933462797 0.000000000 0.000000000
>>>
>>> 0.000000000 20.000000000 0.000000000
>>>
>>> 0.000000000 0.000000000 20.000000000
>>> ---------------------------------------------------------------
>>>
>>>
>>> Thanks
>>>
>>> Rameswar Bhattacharjee
>>> Georgetown University
>>>
>>> On Wed, Jan 11, 2023 at 3:21 AM Giuseppe Mattioli <
>>> giuseppe.mattioli at ism.cnr.it> wrote:
>>>
>>>>
>>>> Dear Rameswar Bhattacharjee
>>>> I don't know if this is the actual source of error, but I strongly
>>>> recommend to use norm-conserving pseudopotentials when performing EXX
>>>> calculations. You have generally no speed-up with US because ecutrho
>>>> is more or less often the same, and in your case you are using
>>>> tremendously high (and likely useless) values for ecutwfc and ecutrho.
>>>> Moreover, using NC you can finely tune the density basis set used to
>>>> build Fock integrals with ecutfock, otherwise stuck to the ecutrho
>>>> value. This combination of values is probably responsible for a very
>>>> high memory request (the code prints an estimate at the beginning of
>>>> the output). For example, *using C and H ONCV pseudopotentials*, you
>>>> should be on the reasonably accurate side using
>>>>
>>>> ecutwfc=90.0
>>>> ecutfock=135.0~180.0
>>>> ecutrho is automatically fixed at ecutwfc*4=360.0 in the case of NC
>>>>
>>>> HTH
>>>> Giuseppe
>>>>
>>>> Quoting Rameswar Bhattacharjee <rb1820 at georgetown.edu>:
>>>>
>>>>> Hi Everyone,
>>>>> I am trying to get a band structure using a hybrid functional. As
>>>> suggested
>>>>> previously in this forum I choose PBE0 functional with pbe-ultrasoft
>>>>> pseudopotential. The calculation starts fine and scf iteration converges
>>>>> smoothly. But the job is stuck and does not complete with the last line
>>>> "ACE
>>>>> projected onto 60 (nbndproj) and applied to 60 (nbnd) bands". I am
>>>>> not sure what mistake I am making here or if it is just a memory issue as
>>>>> no error message is printing. Any suggestion would be highly
>>>> appreciated. I
>>>>> am providing last few lines of the output and my input str for reference.
>>>>> Thank you.
>>>>>
>>>>> Last few lines of the scf output
>>>>> -------------------------------
>>>>>
>>>>> highest occupied, lowest unoccupied level (ev): -3.5627 -3.2294
>>>>>
>>>>>
>>>>> ! total energy = -312.15802972 Ry
>>>>>
>>>>> estimated scf accuracy < 3.2E-09 Ry
>>>>>
>>>>>
>>>>> convergence has been achieved in 12 iterations
>>>>>
>>>>>
>>>>> Using ACE for calculation of exact exchange
>>>>>
>>>>>
>>>>> EXX grid: 4188907 G-vectors FFT dimensions: ( 100, 288, 288)
>>>>>
>>>>> ACE projected onto 60 (nbndproj) and applied to 60 (nbnd)
>>>> bands
>>>>> ----------------------------------
>>>>>
>>>>> Input str
>>>>> ------------------------------
>>>>>
>>>>> &control
>>>>>
>>>>> restart_mode='from_scratch'
>>>>>
>>>>> calculation='scf'
>>>>>
>>>>> prefix='Test',
>>>>>
>>>>> pseudo_dir = '/home/QE-PP',
>>>>>
>>>>> outdir='./scratch'
>>>>>
>>>>> /
>>>>>
>>>>> &system
>>>>>
>>>>> ibrav = 0,
>>>>>
>>>>> tot_charge=0,
>>>>>
>>>>> nat=36,
>>>>>
>>>>> ntyp=2,
>>>>>
>>>>> ecutwfc = 140,
>>>>>
>>>>> ecutrho = 1400.0,
>>>>>
>>>>> input_dft='pbe0'
>>>>>
>>>>> vdw_corr='grimme-d2',
>>>>>
>>>>> nbnd = 60
>>>>>
>>>>> /
>>>>>
>>>>> &electrons
>>>>>
>>>>> electron_maxstep = 1000
>>>>>
>>>>> mixing_beta=0.7
>>>>>
>>>>> diagonalization='david'
>>>>>
>>>>> diago_david_ndim = 2
>>>>>
>>>>> conv_thr=1.0d-8
>>>>>
>>>>> /
>>>>>
>>>>> &ions
>>>>>
>>>>> /
>>>>>
>>>>> &CELL
>>>>>
>>>>> cell_dynamics='bfgs',
>>>>>
>>>>> press = 0.0
>>>>>
>>>>> cell_dofree = 'x',
>>>>>
>>>>> /
>>>>>
>>>>> ATOMIC_SPECIES
>>>>>
>>>>> H 1.008 H.pbe-rrkjus_psl.1.0.0.UPF
>>>>>
>>>>> C 12.0107 C.pbe-n-rrkjus_psl.1.0.0.UPF
>>>>>
>>>>>
>>>>> ATOMIC_POSITIONS {angstrom}
>>>>>
>>>>>
>>>>> K_POINTS automatic
>>>>>
>>>>> 12 1 1 0 0 0
>>>>>
>>>>>
>>>>> CELL_PARAMETERS {angstrom}
>>>>>
>>>>> 6.933462797 0.000000000 0.000000000
>>>>>
>>>>> 0.000000000 20.000000000 0.000000000
>>>>>
>>>>> 0.000000000 0.000000000 20.000000000
>>>>>
>>>>> ----------------
>>>>>
>>>>>
>>>>> Rameswar Bhattacharjee
>>>>>
>>>>> Georgetown University
>>>>
>>>>
>>>>
>>>> GIUSEPPE MATTIOLI
>>>> CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
>>>> Via Salaria Km 29,300 - C.P. 10
>>>> I-00015 - Monterotondo Scalo (RM)
>>>> Mob (*preferred*) +39 373 7305625
>>>> Tel + 39 06 90672342 - Fax +39 06 90672316
>>>> E-mail: <giuseppe.mattioli at ism.cnr.it>
>>>>
>>>> _______________________________________________
>>>> The Quantum ESPRESSO community stands by the Ukrainian
>>>> people and expresses its concerns about the devastating
>>>> effects that the Russian military offensive has on their
>>>> country and on the free and peaceful scientific, cultural,
>>>> and economic cooperation amongst peoples
>>>> _______________________________________________
>>>> Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
>>>> users mailing list users at lists.quantum-espresso.org
>>>> https://lists.quantum-espresso.org/mailman/listinfo/users
>>>>
>>
>>
>>
>> GIUSEPPE MATTIOLI
>> CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
>> Via Salaria Km 29,300 - C.P. 10
>> I-00015 - Monterotondo Scalo (RM)
>> Mob (*preferred*) +39 373 7305625
>> Tel + 39 06 90672342 - Fax +39 06 90672316
>> E-mail: <giuseppe.mattioli at ism.cnr.it>
>>
>> _______________________________________________
>> The Quantum ESPRESSO community stands by the Ukrainian
>> people and expresses its concerns about the devastating
>> effects that the Russian military offensive has on their
>> country and on the free and peaceful scientific, cultural,
>> and economic cooperation amongst peoples
>> _______________________________________________
>> Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
>> users mailing list users at lists.quantum-espresso.org
>> https://lists.quantum-espresso.org/mailman/listinfo/users
>
>
> _______________________________________________
> The Quantum ESPRESSO community stands by the Ukrainian
> people and expresses its concerns about the devastating
> effects that the Russian military offensive has on their
> country and on the free and peaceful scientific, cultural,
> and economic cooperation amongst peoples
> _______________________________________________
> Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
> users mailing list users at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users
GIUSEPPE MATTIOLI
CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
Via Salaria Km 29,300 - C.P. 10
I-00015 - Monterotondo Scalo (RM)
Mob (*preferred*) +39 373 7305625
Tel + 39 06 90672342 - Fax +39 06 90672316
E-mail: <giuseppe.mattioli at ism.cnr.it>
More information about the users
mailing list