ABINIT Discussion Forums

Greetings all,

I have been using Quantum Espresso for a few years now and I am atttempting to learn abinit through the tutorials as well as reproducing previous results from QE. So far, I have succeeded in all but one task, which is reproducing the band structure of graphene. I start with a configuration for the rprim, acell , and xangst for the atomic positions. I have calculated multiple datasets for changing acell to find the minimum energy and ionmov=3 for relaxing the atomic positions. The energy minimization curves look good in that I get well-fitted data, but when I try to run the band structure along high symmetry lines, I still get a band gap at K. Of course, I am sure this is due to my mistakes, so I am hoping that someone can point out where I have gone wrong. For brevity, I am attaching one input file which shows the calculation parameters for the band structure which are a result of the convergence studies.

Thanks for any help
Jeff Mullen
North Carolina State University

Dear Jeff,

The coordinates of K point is incorrect, which should be (0.577639, 0.577639, 0) in your system. By the way, an ecut of 120 Ha is too large for carbon; the comment for xangst is incorrect.

Sincerely,
Guangfu Luo

Thanks for your response.

I replaced my k points in the input file as you suggested and the final energies calculated are

kpt= 0.5776 0.5776 0.0000 (reduced coord)
-14.42183 -13.59451 -8.77891 -2.58747 0.75599 9.20998 10.94101 11.02132

I am unsure where to go from here. Either my geometry is wrong, or I am calculating the wrong k point path (using the wrong coordinates).

Thanks for your help
Jeff Mullen
North Carolina State University

P.S. Attached are my new input file and the .files (named is changed because I could not upload the "files" extension)

Dear Jeff,

Check the files I use and the results they produce.

Sincerely,
Guangfu Luo

Thanks for your help. Between your input file and several hours of experimentation I fixed the multiple problems I had with my geometry and am now reproducing the desired results.

Jeff Mullen
North Carolina State University

Hello,


in passing, you shouldn't need 120 Ha unless you are doing all electron carbon (note-pwscf uses Ry)

your atomic positions were probably not precise enough (use reduced coordinates 1/3 2/3 0 to make sure you preserve symmetry)

The K point, however, was correct as stated "1/3 1/3 0"

You could have problems because the default shift for k-points (0.5 0.5 0.5) has not been removed. You might check this and reset it with "shiftk 0 0 0"

Matthieu

Hi Matthieu,

Thanks for your comments. Since making the original post, I have improved my understanding of the units and how to declare a unit to be in eV instead of Hartree when I need it as well as many other parameters. Once I realized these mistakes, I immediately calculated the correct results.

Regards
Jeff