3-electrode example in TranSIESTA

Asked by Zhenfei Liu on 2016-10-25


I am wondering if Nick could provide a complete set of example input files for a 3-electrode calculation, such as something like the T-junction in Fig. 1b of this paper (http://dx.doi.org/10.1016/j.carbon.2016.01.084). I know how to do a 2-electrode calculation, and want to make sure my 3-electrode system is setup correctly, especially the relative coordinates of the 3rd electrode and the box size of the scattering region calculation. I am particularly confused by how to understand the periodicity along the "transverse" direction if there is a 3rd electrode. Since this feature is new in this version, and I haven't found such example files anywhere.

Thanks a lot!

Question information

English Edit question
Siesta Edit question
No assignee Edit question
Solved by:
Nick Papior
Last query:
Last reply:
Nick Papior (nickpapior) said : #1

There are examples in the Tests/TranSiesta-TBTrans folder with 3 and 4 electrodes.

Please consult these examples for directions on how to setup other N-electrode systems.

Zhenfei Liu (zfliu) said : #2

Thank you Nick, I did not realize that there are examples in Tests/.

However I was thinking about something more complex than the examples there so I want to further confirm with you. In the folders you mentioned, both systems are "flat" in the sense that all the y coordinates of all atoms are the same so your box along y direction can be just a number with sufficient vacuum. And if I think of the current goes along z direction (I know in more than 2 electrode setup this is not the only direction for current flow), then the electrode atoms does not span the whole x-y plane (in the sense that there is vacuum between "lines" of electrodes, where the "lines" are along z).

My question is, suppose you have Au(111) surface, say 4x4 atoms on each layer. And you start with the regular two electrode setup for a junction, say, the standard Au-BDT-Au. In this case, my box is about 10A x 10A x 50A big and the "transport direction" is along z and my electrode spans the whole x-y plane without any vacuum (different from the examples). Now what happens if you want to insert a 3rd electrode which is parallel to x-y? If my 3rd electrode is still Au(111) and, say, I need 3 layers of that and another additional a few layers to screen the interaction, so the x-y dimension (10A x 10A) seems to be not enough to accommodate the length required for the 3rd electrode, because that would be about 15A along x (or y), perpendicular to z. In this case, am I enforced to use a larger unit cell, say the 6x6 or even larger dimension along x and y?

I am confused by this. Hope my question is clear. Thank you very much for your time!

Best Nick Papior (nickpapior) said : #3

Yes you are forced to use a bigger unit cell. If your atoms doesn't fit in
the box, you HAVE to increase the box size.

Secondly in such a setup you cannot make periodic boundary conditions for
the gold electrodes in both x and y directions. You will have to make
sacrifices on the electrode bulk electronic structure. Whether you will
retain one periodic direction or none is up to you. But you should of
course test whether the electronic structure is adequate.

Zhenfei Liu (zfliu) said : #4

Thanks Nick, this is really helpful.

One more question regarding your statement "Secondly in such a setup you cannot make periodic boundary conditions for
the gold electrodes in both x and y directions."

 Is this because the gold electrode that is periodic in both x and y will necessarily interact (artificially) with the 3rd electrode which is along either x or y direction? And the 3rd electrode cannot be periodic in x-z or y-z either, as far as I understand.

Nick Papior (nickpapior) said : #5

Yes. :)

Zhenfei Liu (zfliu) said : #6

Thanks Nick Papior, that solved my question.