The ABINIT Users Mailing List ( CLOSED )

[Xavier Gonze <gonze@pcpm.ucl.ac.be>]

Dear David,

Setting stmbias to a non-zero value activates a modification
of the occupation numbers needed to build the
density, see lines 243-253
of newocc.F90, corresponding to the difference between the
normal occupation numbers, and a shifted (by stmbias) occupation  
function -
in effect selecting an energy windows.
That's all ...

One example is given, Tv4#46, with 3 datasets, the first for the
self-consistent computation of the density, and the
two next for nonzero stmbias calculations, non-zero getwfk,
nstep 1 and iscf 5 (so not the usual choice for non-self-consistent,  
but in effect doing it).

This is not what you would like to do.
Actually, I think that you have to experiment a bit.
You might try to use getden 1, with iscf -3 to see whether you can use
a finer grid, for the non-self-consistent part of the computation only,
to get the accuracy that you want ...

Xavier


On 31 Oct 2007, at 20:28, David Pullman wrote:

> Hi,
> Thanks for your response.  I think I'm still unsure as to how the  
> command "stmbias" actually works.  My impression of how one would  
> generate STM image data in ABINIT is (please correct me if / where  
> I'm wrong): 1) do a self-consistent energy calculation for a slab  
> using a k-point grid sufficiently dense to converge the energy.  
> Although this grid may be dense enough to converge the energy, it  
> may be too sparse in the region near the Fermi level to generate a  
> good STM image 2) So, one then does a non-self-consistent energy  
> calculation with many k-points very near the Fermi level.  I had  
> thought that "stmbias" implements step 2, but perhaps I'm wrong?
> Dave
>
> At 11:14 PM 10/30/2007, you wrote:
> > Hi,
> >
> > From the informations in the "input variable" files, it looks like  
> > all
> > the k-point are used. You just build a new density using all k-points
> > and wave functions close to the Fermi level.
> > If your k-mesh is converged this looks sensible. Otherwise I would  
> > say
> > you'd better restart your ground state calculation with a better  
> > mesh.
> >
> > Regards
> >
> > PMA
> >
> > On 10/30/07, dpullman@sciences.sdsu.edu  
> > <dpullman@sciences.sdsu.edu> wrote:
> > > Dear colleagues,
> > >
> > > When ABINIT computes STM data using the command "stmbias energy- 
> > range", how many k-points are sampled near the Fermi level? I  
> > would guess that it's more than ABINIT used in the previous self- 
> > consistent calculations of the energies and wavefunctions, but I  
> > can't seem to find this information in the output files.  Also, is  
> > there a way of selecting which k-points are actually sampled by  
> > the stmbias routine?
> > >
> > > Thanks in advance
> > > Dave Pullman
> > >
> >
> >
> > --
> > Pierre-Matthieu Anglade