[Pw_forum] hybrid functional

"Valentina Dellacà C.R.F. S.C.p.A." valentina.dellaca at tirocinanti.crf.it
Wed Apr 10 11:13:25 CEST 2013


Hi, thanks for the help!
Valentina

Il 04/10/2013 10:24 AM, mohnish pandey ha scritto:
> Hi Valentina!
>
> Yea. You described it right :)
>
>
> On Wed, Apr 10, 2013 at 9:07 AM, "Valentina Dellacà C.R.F. S.C.p.A." 
> <valentina.dellaca at tirocinanti.crf.it 
> <mailto:valentina.dellaca at tirocinanti.crf.it>> wrote:
>
>     Dear  Monhish,
>     thanks again for your answer. So, if I am understanding correctly
>     I should just perform a scf run with a dense k-point grid and then
>     go on with the DOS using projwfc.x or dos.x, that's right? I was
>     performing a nscf run after the scf one because I was following
>     the tutorial:
>
>     http://www.fisica.uniud.it/~giannozz/QE-Tutorial/
>     <http://www.fisica.uniud.it/%7Egiannozz/QE-Tutorial/>
>     *Post-processing and data analysis* (slides
>     <http://www.fisica.uniud.it/%7Egiannozz/QE-Tutorial/tutorial_postproc.pdf>,
>     examples
>     <http://www.fisica.uniud.it/%7Egiannozz/QE-Tutorial/examples_postproc.tgz>)
>
>     Regards,
>     Valentina
>
>
>     Il 04/09/2013 04:20 PM, mohnish pandey ha scritto:
>>     Dear Valentina,
>>
>>     You can do the calculations with the SCF run in HSE/PBE0, that
>>     shouldn't be a problem. And why do you need nscf calculation
>>     anyway to do the DOS calculation. 'nscf' calculation is needed
>>     when you are interested in band-structure where you use the
>>     ground state charge density for interpolation along certain
>>     directions. Your scf calculations will won't have any problem in
>>     calculation the DOS. Hope it helps :)
>>
>>
>>     On Tue, Apr 9, 2013 at 3:10 PM, "Valentina Dellacà C.R.F.
>>     S.C.p.A." <valentina.dellaca at tirocinanti.crf.it
>>     <mailto:valentina.dellaca at tirocinanti.crf.it>> wrote:
>>
>>         Dear Arles,
>>         thanks for your answer. My doubts are about the employment of
>>         the projwfc.x, post processing tool, when using hybrid
>>         functional,  since the README of the hybrid functional
>>         tutorial says that one gets into troubles when performing
>>         non-scf calculations. At the moment I see the way to proceed
>>         if I wish to plot the band structure from the log file of the
>>         scf run, but i do not know how to get informations about the
>>         dos from such file, since it (as far as I understand) seems
>>         to be the only run i can perform when employing hybrid
>>         functionals. Do you know how to get informations about the
>>         dos from the .log of a scf run?
>>         Thank you!!!
>>         Valentina
>>
>>         Il 04/09/2013 02:24 PM, Arles V. Gil Rebaza ha scritto:
>>>         Dear Valentina,
>>>         Yes, projwfc.x is a postprocessing tool like dos.x, whether
>>>         you want to get only the density of state, "dos.x" is
>>>         enough, whereas "projwfc.x" you can obtain DOS, projected
>>>         DOS (PDOS) and the magnetic moment per atom too.
>>>         http://www.quantum-espresso.org/?page_id=876
>>>
>>>         Best
>>>
>>>         PhD Arles V. Gil Rebaza
>>>         Instituto de Física de La Plata
>>>         Argentina
>>>
>>>
>>>         2013/4/9 "Valentina Dellacà C.R.F. S.C.p.A."
>>>         <valentina.dellaca at tirocinanti.crf.it
>>>         <mailto:valentina.dellaca at tirocinanti.crf.it>>
>>>
>>>             Dear Mohnish,
>>>             thanks for you answer. I will follow your suggestion and
>>>             try also HSE, with a coarser k-point mesh, (indeed with
>>>             PBE0 the computation time is quiet large), and I did not
>>>             know about projwfc.x.
>>>             I have a further question concerning the DOS. If I wish
>>>             to know about the DOS of my system, which uses  the
>>>             hybrid functional implementation, am I able to use the
>>>             projwfc.x? Is'n it a postprocessing tool?
>>>             Thanks,
>>>             regards,
>>>             Valentina
>>>
>>>             Il 04/09/2013 11:18 AM, mohnish pandey ha scritto:
>>>>             Hi,
>>>>
>>>>             As far as I can understand there is no rule. It depends
>>>>             how fine you want to represent your Fock operator.
>>>>             Normally calculations of the Fock exchange in
>>>>             expensive, so its not affordable to do very fine grid
>>>>             calculations  So I suggest you start with low nqx(1, 2,
>>>>             3) and increase it systematically. Increasing it will
>>>>             give you better results( I mean converged results, not
>>>>             necessarily close to experiments).
>>>>
>>>>             Just an advice, PBE0 needs much more high density
>>>>             k-point mesh for the convergence of Fock exchange since
>>>>             it's a long range interaction. So you have to do
>>>>             calculations with high density of k-points. But if you
>>>>             use HSE instead of PBE0 you get away with coarser
>>>>             k-point mesh since it involves screened exchange
>>>>             interaction which converges much faster than bare
>>>>             Coulomb exchange as in PBE0. Regarding to the DOS plot
>>>>             you will find a nice tutorial on Heither's website, I
>>>>             am pasting the link
>>>>             http://www.stanford.edu/~hkulik/www/Tutorials/Entries/2012/1/24_Population_analysis_and_PDOS.html
>>>>             <http://www.stanford.edu/%7Ehkulik/www/Tutorials/Entries/2012/1/24_Population_analysis_and_PDOS.html>
>>>>
>>>>
>>>>             On Tue, Apr 9, 2013 at 10:05 AM, "Valentina Dellacà
>>>>             C.R.F. S.C.p.A." <valentina.dellaca at tirocinanti.crf.it
>>>>             <mailto:valentina.dellaca at tirocinanti.crf.it>> wrote:
>>>>
>>>>                 Hi,
>>>>                 I am a Quantum Espresso beginner user, and I am
>>>>                 simulating the Ta2O5 structure, with the employment
>>>>                 of the hybrid functional PBE0. I have some questions :
>>>>
>>>>                 - if I choose a k-point grid not in automatic, how
>>>>                 should I pick the nqx1,2,3? Is there a rule?
>>>>                 - when one wishes to have informations (ex. plot)
>>>>                 about the density of states from the scf run, how
>>>>                 should I proceed?
>>>>
>>>>                 Thank you,
>>>>                 regards
>>>>                 Valentina
>>>>>
>>>>>                 -- 
>>>>
>>>>
>>>>                 -- 
>>>>
>>>>                 _______________________________________________
>>>>                 Pw_forum mailing list
>>>>                 Pw_forum at pwscf.org <mailto:Pw_forum at pwscf.org>
>>>>                 http://pwscf.org/mailman/listinfo/pw_forum
>>>>
>>>>
>>>>
>>>>
>>>>             -- 
>>>>             Regards,
>>>>             MOHNISH,
>>>>             -----------------------------------------------------------------
>>>>             Mohnish Pandey,
>>>>             PhD Student,
>>>>             Center for Atomic Scale Materials Design,
>>>>             Department of Physics,
>>>>             Technical University of Denmark
>>>>             -----------------------------------------------------------------
>>>
>>>
>>>             -- 
>>>
>>>             _______________________________________________
>>>             Pw_forum mailing list
>>>             Pw_forum at pwscf.org <mailto:Pw_forum at pwscf.org>
>>>             http://pwscf.org/mailman/listinfo/pw_forum
>>>
>>>
>>>
>>>
>>>         -- 
>>>         ###--------->   Arles V. <---------###
>>
>>
>>         -- 
>>
>>         _______________________________________________
>>         Pw_forum mailing list
>>         Pw_forum at pwscf.org <mailto:Pw_forum at pwscf.org>
>>         http://pwscf.org/mailman/listinfo/pw_forum
>>
>>
>>
>>
>>     -- 
>>     Regards,
>>     MOHNISH,
>>     -----------------------------------------------------------------
>>     Mohnish Pandey,
>>     PhD Student,
>>     Center for Atomic Scale Materials Design,
>>     Department of Physics,
>>     Technical University of Denmark
>>     -----------------------------------------------------------------
>
>
>     -- 
>
>
>
>
> -- 
> Regards,
> MOHNISH,
> -----------------------------------------------------------------
> Mohnish Pandey,
> PhD Student,
> Center for Atomic Scale Materials Design,
> Department of Physics,
> Technical University of Denmark
> -----------------------------------------------------------------


-- 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20130410/19775e30/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: firma 2013.jpg
Type: image/jpeg
Size: 19579 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20130410/19775e30/attachment.jpg>


More information about the users mailing list