[DFTB-Plus-User] NEGF_MPI: rare DOS results

Gabriele Penazzi penazzi at uni-bremen.de
Fri Nov 6 17:50:29 CET 2015


Hi Matias,

First a point about the input. When you run a non-scc calculation the
code should actually complain if you specify any Eigensolver other than
TransportOnly. The Eigensolver block takes care of the solution of the
density, and if SCC=No is specified, the code should refuse to continue
to perform instructions in the Analysis block if not explicitely
instructed (hence the "dummy" Eigensolver TransportOnly).

With this input file the code should exit with an error code, doesn't it?

Second, I maybe you refer to the "small" spikes which look like noise,
as the singularities are normal.

I guess that your system is a zigzag. Sometimes small "numerical"
scattering can be to small distorsion in the lattice.

However, assuming that you can rule this out, there's still numerical
issue which can prevent an "ideal" solution. I think that the decimation
algorithm used to evaluate the self energies struggle a bit with band
crossing in Gamma, it happened to me as well to see a small spike at
Fermi level in a zigzag tube (never in armchair, though). Usually they
are in the order of 1meV width therefore I didn't really bother (how
many k-points would you need to have a resolution reliable up to 1
meV?). You can reduce their width by decreasing Delta (see manual), but
eventually you'll hit anyway some hardcoded parameters in the decimation
implementation.

Still, I can't magnify clearly the picture and I don't understand if
they should be matter of concern. What's the width of the spikes? Did
you generate the structure numerically by repetition of a unit cell or
you relaxed it?

Best,
Gabriele


On 11/06/2015 04:18 PM, matias berdakin wrote:
> sorry I forgot to say bye in the last mail
> 
> Thanks in advance 
> Matias Berdakin
> 
> 2015-11-06 10:18 GMT-03:00 matias berdakin <matiasberdakin at gmail.com
> <mailto:matiasberdakin at gmail.com>>:
> 
>     Hi everybody
> 
>     I'm doing my firsts trials with the DFTB+_NEGF_MPI (after the
>     Tutorial examples). I'm computing the DOS and the transmission
>     function (non scc) of a conducting carbon nanotube. At  first
>     glance, the overall feature of the DOS is what is expected for this
>     system (see attached figure). Nevertheless  the result present a
>     series of symmetric spikes that make no sense.
>     I can not find the error in the calculation. So an idea is welcome. 
>     At the end of the message I have pasted the geometry and the
>     dftb_in.hsd files.
>     Imágenes integradas 1
>     Geometry = GenFormat {
>     <<< 'coords_grande.gen'
>     }
> 
>     Transport {
>       Device {
>         AtomRange = 1 240
>         FirstLayerAtoms =  1 121
>       }
>       Contact {
>         Id = "source"
>         AtomRange = 241 480
>         FermiLevel [eV] = -4.602
>         potential [eV] = 0.0
>       }
>       Contact {
>         Id = "drain"
>         AtomRange = 481 720
>         FermiLevel [eV] = -4.602
>         potential [eV] = 0.0
>       }
>     }
> 
>     Hamiltonian = DFTB {
>       SCC = No
>       MaxAngularMomentum = {
>         C = "p"
>       }
> 
>       SlaterKosterFiles = Type2FileNames {
>         Prefix = "/mio-1-1/" 
>                                 
>         Separator = "-"
>         Suffix = ".skf"
>       }
> 
>     Eigensolver = GreensFunction {
>     ContourPoints = 50 50
>     }
> 
>     Analysis = {
>         TunnelingAndDos {
>           Verbosity = 101
>           EnergyRange [eV] = -7.0 -3.0
>           EnergyStep [eV] = 0.005
>           Region = {
>             Atoms = 1 240
>           }
>         }
>     }
>             
> 
> 
> 
> 
> _______________________________________________
> DFTB-Plus-User mailing list
> DFTB-Plus-User at mailman.zfn.uni-bremen.de
> https://mailman.zfn.uni-bremen.de/cgi-bin/mailman/listinfo/dftb-plus-user
> 

-- 
--
Dr. Gabriele Penazzi
BCCMS - University of Bremen

http://www.bccms.uni-bremen.de/
http://sites.google.com/site/gabrielepenazzi/
phone: +49 (0) 421 218 62337
mobile: +49 (0) 151 19650383


More information about the DFTB-Plus-User mailing list