The ABINIT Users Mailing List ( CLOSED )

[verstraete@pcpm.ucl.ac.be]

A few details on occupations: I think you've used occopt 1 (default) which
means semiconductor occupations. In this regime the Fermi level isn't well
defined, so bands are filled with occupation 2 until there are no
electrons left. This means even if you have a "gap" of 0.00001 eV, the
occupations will be 2 2 2 2 0.0 0.0 0.0

ZnSe is a dielectric, so I suspect this is correct: simply the Fermi
energy is not defined inside abinit, it's not needed. You can suppose it's
at half the gap. Otherwise, you should be using a metallic occupation
scheme (occopt >= 3). In these cases the distribution can be FD, gaussian,
or cold smearing (see ABINIT variable pages), and the Fermi energy will be
defined by conservation of the number of electrons. In these cases only
does ABINIT print out a line with

"newocc : new Fermi energy is"  to the log file for every SCF iteration,

and at the end of the SCF run:

"Fermi energy (hartree) =" to the output

When in doubt, always use a smearing, eventually with a very small tsmear.
0.001 Ha almost never affects any properties, and gaussian smearing
(occopt 7) is the best controlled one. Cold smearing is the most efficient
(leaves the energy unchanged with T to order 3) but the occupation
function is not monotonous and goes a little bit above 2 - this may be
unethical for some people.

Note that even if you only want the energies around k=0 you still need 
more kpoints to converge the density properly, or a huge cell...

Matthieu


On Mon, 4 Nov 2002, Kristopher Andersen wrote:

> I'm puzzled by the Fermi energy in ABINIT. I've been running some
> calculations on ZnSe and for the band energies at k = 0 I get
> 
>  kpt# 1, nband= 10, wtk=  0.00800, kpt=  0.0000  0.0000  0.0000
>   -0.53351  -0.29563  -0.29563        -0.29563  -0.28408  
>   -0.28408  -0.05693  -0.05693  -0.05693  -0.01967
> 
> The Fermi energy isn't printed out anywhere in the output file, but is
> supposedly written to the unformatted _WFK and _DEN files. When I look
> there, they both give the Fermi energy (variable fermie) to be 0.0?
> Clearly this would be in the conduction band given the band energies
> above, since ZnSe has 9 occupied bands and is a direct gap
> semiconductor.
> 
> Is it true that ABINIT is predicting ZnSe to be a metal or am I just
> not reading the Fermi energy right? Why isn't the Fermi energy printed
> out in the output file somewhere? Maybe it is and I'm missing it?
> 
> Thanks for the help. For reference, the benign input file for this
> calculation is
> 
> ########################################################################
> # ZnSe.scf
> #
> # Crystalline ZnSe, zincblende crystal structure
> # 
> # Goes through the self-consistent cycle and outputs the ground state
> # charge density and wavefunctions.
> ########################################################################
> 
> ixc  11                  # PBE GGA functional
> ecut 45.0                # planewave cutoff energy (Hartree)
> 
> # k-point grid
> kptopt  1                # automatic generation, using symmetry
> nshiftk 1
> shiftk  0.0 0.0 0.0
> ngkpt   5 5 5            # Monkhorst-Pack grid (odd to include gamma)
> 
> # unit cell
> acell 3*10.9118          # scaling parameters for primitive vectors
> rprim  0.0  0.5  0.5     # FCC primitive vectors
>        0.5  0.0  0.5   
>        0.5  0.5  0.0
> 
> # specify atoms
> ntype 2                  # 2 types of atoms (Zn, Se)
> zatnum 30 34             # type 1 = Zn; type 2 = Se
> natom 2                  # 2 atoms in the unit cell
> type 1 2                 # atom 1 = Zn; atom 2 = Se
> xred                     # atomic positions are given in reduced coordinates
>    0.0  0.0  0.0           # reduced coordinates of atom 1 (Zn)
>    0.25 0.25 0.25          # reduced coordinates of atom 2 (Se)
> 
> # SCF parameters
> nstep  25                # maximum number of SCF cycles
> tolvrs 1.0d-6            # tolerance on the potential (Hartree)
> diemac 12.0              # model dielectric constant (for preconditioning)
> prtden 1                 # output the charge density
> 
> 

-- 
===================================================================
Matthieu Verstraete               mailto:verstraete@pcpm.ucl.ac.be 
PCPM, Boltzmann, pl. Croix du Sud, 1            tel: 010/ 47 86 81
B-1348 Louvain-la-Neuve  Belgium                fax: 010/ 47 34 52