Internal error, cannot braket Ef

[jsttse]

Error in bracketing fermi energy
Roxana Margine <emargine@binghamton.edu>
Wed 3/23/2016 8:52 AM
Dear John, We will be happy to answer, could you please post your query on our forum at http://epw.freefo.org/index.php so that other users can benefit from the discussion. Within the next few weeks we are going to release a new version of the code that works
Tse, John
Wed 3/23/2016 8:48 AM
To:
rmargine@binghamton.edu;
Dear Prof. Margine,

I am a novice user of the EPW code.

I am doing a calculation of a metallic solid at high pressure and the Fermi energy is fairly high.

I encountered an error

Fermi energy coarse grid = 17.747441 eV

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
from efermig : error # 1
internal error, cannot braket Ef
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

stopping ...

I did some web searching and I found that this may be an error with QE 4.3

Please see

http://pw_forum.pwscf.narkive.com/feaaZ ... -in-qe-4-3

By the way, I have succeed ed doing the electron-phonon coupling of the same system using an even older code ... about version 2

Regards

John

[sponce]

Dear John,

This issue happen for metallic system.
You need to manually remove the check that raises the error in the code and recompile.

Alternatively you could use the new version of EPW compatible with the latest QE version.

EPW is currently in the trunk of QE on qe-forge: http://qeforge.qe-forge.org/gf/project/q-e/

However if you are not a QE developer, you will not have access to it.

The next public release of the QE code in April will include EPW.

Best,

Samuel

[jsttse]

Dear Samuel,

Thank you for your reply. I have compiled your new branch but I encountered the same problem in finding the Fermi energy. I have also tried different schemes in scf and nscf calculations to compute the Fermi energy but to no avail.

In the epw.out the correct Fermi energy obtained from nscf was always printed correctly e.g.

in nscf.out

-5.0010 -4.8909 -4.8064 -4.7100 -4.6185 3.3877 7.9144 11.1765
15.2214 19.0969 23.0794 25.9918 28.0813

the Fermi energy is 17.7682 ev

while in epw.out

Finished reading Wann rep data from file


Fermi energy coarse grid = 17.768161 eV

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
from efermig : error # 1
internal error, cannot braket Ef
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

stopping ...

Are there any ways to help to find the correct Fermi energy?

Thanks

John

[carla.verdi]

Hi John

Can you please add some more info on your epw.in, in particular what is the fine k grid you are using, and the degaussw parameter?

Carla

[Arnie]

I have this same problem.
In my epw.in file I have:

Code: Select all

degaussw    = 0.02 ! eV
 
  nk1         = 6
  nk2         = 6
  nk3         = 6
  nq1         = 6
  nq2         = 6
  nq3         = 6
  nqf1         = 20
  nqf2         = 20
  nqf3         = 20
  nkf1         = 20
  nkf2         = 20
  nkf3         = 20

How eliminate this error?

[carla.verdi]

Dear Arnie

In my experience, that error can happen if the fine k grid is not good enough to calculate the Fermi level, however that doesn't seem the case given your input. Another reason I can think of is if the Wannier bands are somehow wrong. May I suggest two things:
- play a bit with the nkf and degaussw parameters (but I don't think this is likely to solve the problem)
- input a fermi level manually by inserting 'efermi_read = .true.' and 'fermi_energy=XXX' (for example the value from the nscf calculation). Then the code won't crash and you can see if the band structure is correct by putting 'band_plot=.true.' (and elecselfen, phonselfen etc =.false. to be fast and just do the band structure).

Best
Carla