[Pw_forum] GaAs: band gap with HSE is too large

[Andrea Ferretti]

Hi Jeronimo,

after a quick look it seems you are not defining the q grid (nqx1, nqx2, nqx3, 
SYSTEM namelist) used in the EXX calculation. Not sure it is still like that, 
but in the past the default value used to be just the gamma point (which may 
explain a very poor result for a bulk system like GaAs)
instead, nqx's should be equal to the nk's used to generate the kpt mesh (you 
can try to use a smaller q grid, but carefully checking the convergence)

BTW: as a second check, the parameter governing the range separation may also 
need to be controlled

take care
Andrea




> Dear all, 
> I'm using HSE to study bulk GaAs, but even with the parameters cited in the 
> literature  (number of k points,
> exx_fraction=0.35, lattice parameters=10.77, ecutwfc=50 Ry, etc) I'm not able 
> to reproduce the bandgap. It
> should give ~1.43 eV.  
> 
> I'm getting a very large value ~ 3 eV.  Even with smaller values of 
> exx_fraction =0.15, the gap is still too
> large. I've tested different norm conserving PBE pseudopotentials, k point 
> convergence (up to 8x8x8), energy
> cutoff convergence, but this discrepancy persists. 
> 
> I'm computing the DOS directly from a scf run, and it gives ok except for the 
> value of gap. I would be very
> grateful if someone has any idea or information on this. Cheers, 
> 
> Jeronimo.
> 
> -------------------------
> 
> A typical input file is:
> 
>  &control
>     calculation='scf',
>     restart_mode='from_scratch',
>     prefix='GaAsbulk'
>     pseudo_dir = '/home/j/Pseudos-QE',
>     outdir='./temp',
>     disk_io='high'
>  /
>  &system
>     ibrav=2,
>     celldm(1)=10.77,
>     nat=2,
>     ntyp=2,
>     ecutwfc=60.0d0,
>     nspin=1,
>     occupations='tetrahedra',
>     input_dft='hse',
>     exx_fraction=0.35
>  /
> 
>   &electrons
>     conv_thr=1d-8,
>     mixing_beta=0.7,
>     diagonalization='david'
>  /
> 
>   ATOMIC_SPECIES
>    Ga   69.723   Ga.pbe-hgh.UPF
>    As   74.92160 As.pbe-hgh.UPF
> 
>   ATOMIC_POSITIONS  alat
>     Ga 0.0   0.0   0.0
>     As 0.25  0.25  0.25
> 
>   K_POINTS   automatic
>      5 5 5 0 0 0 
> 
>

-- 
Andrea Ferretti
S3 Center, Istituto Nanoscienze, CNR
via Campi 213/A, 41125, Modena, Italy
Tel: +39 059 2055322;  Skype: andrea_ferretti
URL: http://www.nano.cnr.it