Multi-grid boundary conditions setting

[Bertram Wang]

I have known that boundary conditions are related to Hartree potential. Now, I want to calculate the transimission spectrum of a gate-voltage controled Bilayer graphene nanoribbon, the picture of  which is attached below. In this real case, is there someone could tell me what is the most correct setting? What's the difference between periodic, dirichlet and neumann?
    Thank you!

[kstokbro]

There are several issues you should be aware of:
1. Your electrodes looks too small make them longer ~7 Å
2. make the gate and dielectric such that there is a fair distance from their edges to the central cell boundaries, I would recommend ~10 Å, but less may also work.
3. use neumann boundary conditions in x,y, this will make it possible for the potential to float. However, it also means that the zero bias gate is not well defined, i.e. only relative gate potentials are well defined. To relate the gate potential to zero vacuum, you need to have a dirichlet boundary condition at some side, forinstance the side below the gate.

[Bertram Wang]

Thanks for your help!
1. Do you mean that I should make the dieletric layer more closed(less than 10A)  to the upper plane of GNR?
2. Do you mean that I should use dirichlet boundary for zero gate voltage and neumann boudary for any voltage but zero?