[Pw_forum] projwfc problem: sum of pdos doesn't match total dos

J. D. Burton jlz101 at unlserve.unl.edu
Wed Aug 20 18:34:16 CEST 2008


Hi Derek,

Interesting, though it shouldn't be that difficult to just take the
pseudopotential (PSP) as it is and use the ld1.x program (comes with the
espresso package) and calculate other atomic pseudowavefunctions that you
can then project onto. You can then include these in your PSP file.
I suppose if you're really ambitious, you could alter the projwfc program to
not use the atomic functions in the PSP file, but instead use different
localized functions in a file that you provide (for example, you could try
Wannier functions). I know that you can do this with the lda+u feature, like
setting U_projection_type to 'file'. See the INPUT_PW.txt file for details
on that.

Cheers,
J. D.


************************************
J. D. Burton, Ph.D.
jlz101 at unlserve.unl.edu
University of Nebraska Lincoln
Physics and Astronomy
Office Ph. (402) 472 2499
213C Ferguson Hall
************************************
"The job of a scientist is to generate wrong ideas as fast as possible."
-- Murray Gell-Mann


-----Original Message-----
From: pw_forum-bounces at pwscf.org [mailto:pw_forum-bounces at pwscf.org] On
Behalf Of stewart at cnf.cornell.edu
Sent: Wednesday, August 20, 2008 10:43 AM
To: PWSCF Forum
Subject: Re: [Pw_forum] projwfc problem: sum of pdos doesn't match total dos

Hi J. D.,

Thanks for your comment!  I am currently working on a project where it would

be helpful to know the orbital character of the higher energy dos peaks.  I 
know that they will have little effect on the self-consistent calculation. 

You raise an important point that I perhaps missed with the projwfc 
implementation.  If projwfc can only map back to the orbitals provided in 
the pseudopotential (2s 2p), then this will limit its usefulness for 
determining the orbital character of higher energy conduction bands. 

Thanks, 

Derek 


J. D. Burton writes: 

> Derek, 
> 
> Why would you need to calculate more than 8 bands? (1 s-state + 3
p-states)
> x 2 C atoms per unit cell = 8 bands, and half of them are doubly-occupied.
> By setting nbnd = 20 (10 per atom in the unit cell) you get all the way up
> into d-like-bands, and your pseudopotential file doesn't contain atomic
> d-wavefunctions of carbon. That would explain why the sum of your
projected
> dos doesn't match your total density, i.e. the extra peaks are d-bands. 
> 
> Cheers,
> J. D. 
> 
> ************************************
> J. D. Burton, Ph.D.
> jlz101 at unlserve.unl.edu
> University of Nebraska Lincoln
> Physics and Astronomy
> Office Ph. (402) 472 2499
> 213C Ferguson Hall
> ************************************
> "The job of a scientist is to generate wrong ideas as fast as possible."
> -- Murray Gell-Mann 
> 
> -----Original Message-----
> From: pw_forum-bounces at pwscf.org [mailto:pw_forum-bounces at pwscf.org] On
> Behalf Of stewart at cnf.cornell.edu
> Sent: Wednesday, August 20, 2008 9:39 AM
> To: pw_forum at pwscf.org
> Subject: [Pw_forum] projwfc problem: sum of pdos doesn't match total dos 
> 
> Hi everyone,  
> 
> I am looking at the density of states of graphene and I have run into a 
> problem with the projwfc program.  When I plot out the total density of 
> states for the system and compare this to the sum of the partial density
of 
> states found in the pdos_tot file, I find that the sum of the partial DOS
is 
> 
> missing some large peaks at energies above the Fermi energy.  The total
DOS 
> from the projwfc run matches the total DOS generated using dos.x and the 
> tetrahedra technique.  The bands responsible for these missing peaks also 
> show up in the band plot.  I have tried using both methfessel-paxton and 
> marzari-vanderbilt smearing (degauss 0.01) but this does not have any
effect 
> 
> on the problem.  
> 
> I thought that perhaps the projection onto the atomic orbitals was missing
a 
> 
> significant portion of the charge, however my spilling parameter is fairly

> low 0.0116.  This is comparable to the spilling parameter found in
example08 
> 
> for nickel 0.0115 where the total and partial dos match reasonable well 
> (although slightly shifted in energy in the reference data).  
> 
> I would appreciate any thoughts as to why I am missing these DOS peaks in 
> the partial dos.  Are there difficulties in projecting plane waves to 
> orbitals for conduction bands?  Has anyone else run into this problem?  
> 
> For the details of my calculation, I am using C.pz-rrkjus.UPF, ecutwfc=45,

> ecutrho=450, k-grid for self-consistent calculation (64x64x1), k-grid for 
> non-scf for dos (96x96x1), number of bands=20 and working with espresso 
> 4.0.1.  
> 
> Thanks,  
> 
> Derek
>   
> 
> ################################
> Derek Stewart, Ph. D.
> Scientific Computation Associate
> http://www.people.cornell.edu/pages/das248/
> 250 Duffield Hall
> Cornell Nanoscale Facility (CNF)
> Ithaca, NY 14853
> stewart (at) cnf.cornell.edu
> (607) 255-2856
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum 
> 
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum 
> 
 


################################
Derek Stewart, Ph. D.
Scientific Computation Associate
http://www.people.cornell.edu/pages/das248/
250 Duffield Hall
Cornell Nanoscale Facility (CNF)
Ithaca, NY 14853
stewart (at) cnf.cornell.edu
(607) 255-2856
_______________________________________________
Pw_forum mailing list
Pw_forum at pwscf.org
http://www.democritos.it/mailman/listinfo/pw_forum



More information about the Pw_forum mailing list