[Pw_forum] k-mesh in pwcond

Gabriele Sclauzero sclauzer at sissa.it
Tue Nov 10 09:52:01 CET 2009



Zubaer Hossain wrote:
> Thanks for your replies.
> 
> I know there's no k-mesh input in the input file for pwcond. 

That's not true: after the namelist you must specify a list of 2D k-points, either 
manually (with correct weights) or automatically generated (somehow in the same way as for 
K_POINTS AUTOMATIC in pw.x, but here in 2D instead of 3D).


Actually, I
> was getting different transmission coefficient values with respect to a 
> change in the k-mesh values for the lead and scat calculations. 

If you said that "there's no k-mesh input in the input file for pwcond", how could you see 
those changes in the transmission? Do you mean that you are changing k-points in the pw.x 
calculation, instead?

> 
> I realized that the complex band structures were actually not converged 
> w.r.t the k-mesh. 

This statement does not make much sense to me: k-dependent transmissions need to be 
converged wrt k-mesh, while CBS (which can be computed, for any k_perp point, as a 
function of energy to obtain a k_z VS e_{k_z} plot) has to be converged wrt to ewind and 
epsproj parameters (in order to match with sufficient precision the corresponding bands 
plot from pw.x)


> Seems like pwcond requires quite a dense k-mesh for 
> convergence, especially for metallic systems. However, now I am getting 
> good results.

Which kind of system are you computing? If it is a tipless geometry with 1D periodicity 
(like a monatomic wire with an impurity), you don't need to sample k_perp, since the 
transmission should be k_perp-independent (if you have chosen the cell correctly).
If you are studying a geometry with k-dependent transmission (like tunnel junctions or 
wire between bulk leads), then the convergence of transmission may depend a lot on the 
scattering energy.
Of course you will have metallic leads... what other kind of leads would be of interest to 
you?
I think that the convergence of transmission is related to the CBS of the leads, and in 
particular to the location of band edges in energy.


HTH

GS

> 
> Zubaer
> 
> 
> 
> On Mon, Nov 9, 2009 at 4:20 AM, Gabriele Sclauzero <sclauzer at sissa.it 
> <mailto:sclauzer at sissa.it>> wrote:
> 
> 
> 
>     xirainbow wrote:
>      > Dear :
>      > I do not think pwcond can set k-mesh along the transport direction.
>      > Instead, you can set k-mesh perpendicular  to the transport
>     direction.
>      >
>      > The transport device can be period perpendicular to the transport
>     direction.
>      > However, it is not period along the transport direction
> 
>     This is right, in the pwcond calculation the Born-von Karman PBCs
>     are removed along the
>     transport direction, and you need to take care only of the 2D BZ
>     sampling in the
>     perpendicular directions.
>     Anyway, in the preceding pwscf calculation, periodicity is still
>     there in all 3
>     directions, hence you need a proper k-point sampling also along the
>     transport direction,
>     especially for the leads, which usually are a bulk periodic unit
>     (hence they may not be a
>     supercell).
> 
>     HTH
> 
>     GS
> 
>      >
>      > On Sat, Nov 7, 2009 at 12:46 AM, Zubaer Hossain
>     <zubaexy at gmail.com <mailto:zubaexy at gmail.com>
>      > <mailto:zubaexy at gmail.com <mailto:zubaexy at gmail.com>>> wrote:
>      >
>      >     Dear QE users,
>      >
>      >     Is there any rule for choosing k-meshes for the lead and the
>      >     scattering region along the transport direction?
>      >
>      >     Thanks in advance,
>      >     Zubaer
>      >     UIUC
>      >
>      >
>      >
>      >
>      > --
>      > ____________________________________
>      > Hui Wang
>      > School of physics, Nankai University, Tianjin, China
>      >
>      >
>      >
>     ------------------------------------------------------------------------
>      >
>      > _______________________________________________
>      > Pw_forum mailing list
>      > Pw_forum at pwscf.org <mailto:Pw_forum at pwscf.org>
>      > http://www.democritos.it/mailman/listinfo/pw_forum
> 
>     --
> 
> 
>     o ------------------------------------------------ o
>     | Gabriele Sclauzero, PhD Student                  |
>     | c/o:   SISSA & CNR-INFM Democritos,              |
>     |        via Beirut 2-4, 34014 Trieste (Italy)     |
>     | email: sclauzer at sissa.it <mailto:sclauzer at sissa.it>              
>               |
>     | phone: +39 040 3787 511                          |
>     | skype: gurlonotturno                             |
>     o ------------------------------------------------ o
>     _______________________________________________
>     Pw_forum mailing list
>     Pw_forum at pwscf.org <mailto:Pw_forum at pwscf.org>
>     http://www.democritos.it/mailman/listinfo/pw_forum
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum

-- 


o ------------------------------------------------ o
| Gabriele Sclauzero, PhD Student                  |
| c/o:   SISSA & CNR-INFM Democritos,              |
|        via Beirut 2-4, 34014 Trieste (Italy)     |
| email: sclauzer at sissa.it                         |
| phone: +39 040 3787 511                          |
| skype: gurlonotturno                             |
o ------------------------------------------------ o



More information about the users mailing list