Siesta

About the boundary conditions in transport direction in TRANSIESTA calculations

Asked by Wei

Dear SIESTA developers and users,

I have a few questions about TRANSIESTA:

1, In the TRANSIESTA calculation, is the periodic boundary condition always applied to the transport direction for any two-probe devices, whether they have the same or different electrodes? When we plot the Hartree potential from the .VH file, it seems that periodic condition has been forced for any two-probe systems. Note that I do not mean the relaxation process, but the SCF calculations with NEGF. For devices with the same electrodes, periodic condition can be assumed. However, for devices with different electrodes, it is obviously not reasonable to assume periodic condition in the transport direction, instead, fixed boundary condition and solution in real space should be taken.

2, If we have the option to perform the TRANSIESTA calculation without the need to assume periodic boundary condition in the transport direction for devices with two different electrodes, how shall we set the input parameters?

Thank you very much!
Wei

Question information

Language:
English Edit question
Status:
Solved
For:
Siesta Edit question
Assignee:
No assignee Edit question
Solved by:
Nick Papior
Solved:
Last query:
Last reply:
Revision history for this message
Best Nick Papior (nickpapior) said : 		#1

1. Yes and no (but more no :)).
The FFT (periodic) solution of the Poisson equation is used. However, for a 2 terminal system the boundary conditions of the applied voltage can be corrected by adding a ramp in the unit-cell. Thus the calculated Hartree potential is mathematically correct.

Even for different electrodes this assumption is ok. However, we agree that it would be better to have some combination of Dirichlet and/or Neumann BC depending on the actual setup.

2. Only the FFT solution is implemented in Siesta. So we can't do anything yet.

Revision history for this message
Wei (weix226) said : 		#2

Thanks for the response.

Do you have the plan of considering the Dirichlet and/or Neumann BC in the future versions?

Wei

Revision history for this message
Wei (weix226) said : 		#3

Thanks Nick Papior, that solved my question.