ABINIT Discussion Forums

Dear Abinit users

I am trying to obtain the band gap of a boron doped 2by2 graphene supercell. I used GGA for exchange correlation functional and Troullier-Martins norm conserving pseudopotentials. I obtained the band gap as 0.65893 eV, but the reported value using GGA and norm conserving pseudopotentials implemented in SIESTA package is 0.54 eV.
I don't understand why this much difference in band gap value comes on comparison.
Are the band gap values from different softwares comparable?

Could anyone share some information about this.

Please shed some light on this.

Regards
Seba

are you using exactly the same pseudopotentials as in the Siesta calcs? I don't mean just TM and GGA, I mean really created with all exactly the same input parameters. If not, it's not surprising that the values come out differently. Abinit has been checked quite carefully against other planewave pseudopotentials codes, like quantum espresso, and when exactly the same inputs are used, the outputs are are very very close to each other.

jzwanzig wrote:are you using exactly the same pseudopotentials as in the Siesta calcs? I don't mean just TM and GGA, I mean really created with all exactly the same input parameters. If not, it's not surprising that the values come out differently. Abinit has been checked quite carefully against other planewave pseudopotentials codes, like quantum espresso, and when exactly the same inputs are used, the outputs are are very very close to each other.

Dear Sir

In the Siesta calculation based paper, the computational approach it is mentioned that " Valence electrons have been represented with double zeta basis sets of orbitals localized on atoms; influence of core electrons were covered within pseudopotential formalism. Norm-conserving Troullier-Martins nonlocal pseudopotentials in the Kleinman-Bylander form have been used to account for electron-ion interactions. The energy cut off determining the density of the utilized real space grid was set to 800 Ry. Structural optimization has been conducted employing the conjugate gradient algorithm to achieve residual forces acting on atoms lower than 0.001 eV/Angstrom".

In my calculations, I used GGA in the PBE form to calculate the exchange correlation energy using Troullier-Martins type peusdopotentials. A plane wave basis set with a kinetic energy cutoff energy of 816 eV and structural optimization was conducted using BFGS with maximal absolute force tolerance of 0.0025 eV/Angstrom.

Can I compare the results from Abinit and Siesta based on the above calculation details.

I would also like to know whether electronic bandgap calculation using the above described Abinit computational method is correct or not.

Please shed some light on this.

Thanks and Regards
Seba