[QE-users] Missing empty state in SCF calculation

[Lorenzo Sponza]

Dear everyone,

Thanks for all the suggestions you gave me. Indeed Stefano's suggestion 
solves the issue. If I put "startingwfc = 'random' " the diagonalization 
finds the right bottom conduction even with only one empty state, 
irrespectively of the algorithm used (CG or Davidson).
Instead, If I use the default startingwfc, none of the two 
diagonalization methods catch the right bottom conduction and I need to 
increase the number of empty states to find  it.

How Paolo said, that's not a big deal as long as I need the density, but 
I was intrigued by this unexpected behaviour and I would like to 
understand what's happening.

Best regards


Le 2023-09-01 20:21, Stefano Baroni a écrit :
> You may want to try to randomize the LUMO molecular orbital. SB
> 
> ___
> Stefano Baroni, Trieste -- http://stefano.baroni.me
> 
>> On 1 Sep 2023, at 19:51, Paolo Giannozzi <paolo.giannozzi at uniud.it> 
>> wrote:
>> 
>> On 01/09/2023 18:17, Lorenzo Sponza wrote:
>> 
>>> Dear Paolo, thanks for the answer.
>>> I made a check running the same two calculations using the CG
>>> diagonalization scheme instead of Davidson. The issue persists. May I
>>> send you the input files ?
>> 
>> not a good idea, because I don't think it is a bug and don't have a 
>> solution at hand. You may open an issue on gitlab, though.
>> 
>> Paolo
>> 
>>> Best
>>> Le 2023-09-01 16:49, Paolo Giannozzi a écrit :
>>>> Iterative diagonalization may occasionally miss one or a few states 
>>>> if
>>>> the starting manifold does not contain enough components of the
>>>> missing states. As long as you do not miss any states in the 
>>>> occupied
>>>> manifold, not a big deal for the ground-state calculation (but not 
>>>> for
>>>> the band gap, of course)
>>>> 
>>>> Paolo
>>>> 
>>>> On 9/1/23 16:41, Lorenzo Sponza wrote:
>>>>> 
>>>>> You don't often get email from lorenzo.sponza at onera.fr. Learn why 
>>>>> this
>>>>> is important <https://aka.ms/LearnAboutSenderIdentification>
>>>>> 
>>>>> 
>>>>> Dear all,
>>>>> 
>>>>> I encounter an issue I did not expect and I would like to know why 
>>>>> it
>>>>> happens and how I am suppose to handle it. I'm running an SCF
>>>>> calculation of a gapped system (occupations = 'fixed'). If I 
>>>>> include a
>>>>> large number of empty bands, these are the energies in Gamma:
>>>>> 
>>>>> -4.8200-4.3169-0.9586-0.5478-0.5196-0.4977-0.0537 0.0095 with
>>>>> occupations 1.0000 1.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
>>>>> 
>>>>> Instead, if I include only one empty band the result is
>>>>> 
>>>>> -4.8200-4.3169-0.5477 with occupations 1.0000 1.0000 0.0000
>>>>> 
>>>>> It's like in if the latter case I was missing the state at -0.9586 
>>>>> eV.
>>>>> How it happens ?
>>>>> 
>>>>> I'm rising this issue because I'm checking the gapwidth of a series 
>>>>> of
>>>>> similar systems. I actually thought that it was safe to run simple 
>>>>> SCF
>>>>> calculations with 1 empty state, but I realise that results are
>>>>> possibly wrong. Shall I mandatorily rely on NSCF calculations or 
>>>>> I'm
>>>>> making a mistake somewhere ?
>>>>> 
>>>>> Best regards
>>>>> 
>>>>> -- Dr. Lorenzo Sponza
>>>>> Chargé de Recherche au CNRS
>>>>> Laboratoire d'étude de microstructures (LEM), CNRS-ONERA
>>>>> 29 Avenue de la division Leclerc, 92322 Châtillon
>>>>> Tel: +33146734464
>>>>> 
>>>>> _______________________________________________
>>>>> 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
>>> --
>>> Dr. Lorenzo Sponza
>>> Chargé de Recherche au CNRS
>>> Laboratoire d'étude de microstructures (LEM), CNRS-ONERA
>>> 29 Avenue de la division Leclerc, 92322 Châtillon
>>> Tel: +33146734464
>>> _______________________________________________
>>> 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
>> 
>> --
>> Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
>> Univ. Udine, via delle Scienze 206, 33100 Udine Italy, +39-0432-558216
>> _______________________________________________
>> 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

-- 
Dr. Lorenzo Sponza
Chargé de Recherche au CNRS
Laboratoire d'étude de microstructures (LEM), CNRS-ONERA
29 Avenue de la division Leclerc, 92322 Châtillon
Tel: +33146734464