Vacuum potential

[395235863]

How to calculate the vacuum potential? use the effective potential or electrodifferencepotential?

I think electrostatic potential maybe better.But  ATK cannot calculate it ,isn't it?

I just want to use vacuum potential to do band aligning in the slab model(two material's  vertical interface).

Thank you very much!

[zh]

".But ATK cannot calculate it ,isn't it?". You are right.
If the vacuum thickness is thick enough, both of them can be used to estimate the vacuum potential.
The effective potential, electrodifferenccpotential and electrostatic potential have different convergence dependence on the vacuum thickness.

You can refer to the following paper:
http://iopscience.iop.org/article/10.1088/0953-8984/11/13/006/pdf

[Anders Blom]

The electrostatic potential will be available in ATK 2016.

[395235863]

From the pic we can see ,the effective potential isn't flat in the vacuum region
In my script
vector_c = [0.0, 0.0, 100.0]*Angstrom
boundary_conditions(in the c direction)= [DirichletBoundaryCondition,DirichletBoundaryCondition]]

How to get the reasonable vacuum potential?Must add ghost atom?

Thank you very much!

[Anders Blom]

DirichletBoundaryCondition on both sides is not  a good idea, try Neumann on one.

[395235863]

I tried it. And got the pic. How to get vacuum potential ?(I got the chemical potential)
Thank you!

[Anders Blom]

Since you have now fixed the potential on one side, you know what the energy zero reference is (vacuum). And you know the position of the chemical potential in this reference system, thus you know the distance from the chemical potential to the vacuum level, which is really what you seek.

[395235863]

But in my model , both sides(left and right sides)are vacuum?

the distance from the chemical potential to the vacuum level should be the left one or the right one in the effective potential pic above?or the average value between left and right?

Thank you for your time !

[Anders Blom]

The answer lies in the definition of the Dirichlet vs Neumann boundary condition. One sets the value of the potential, the other the slope. Obviously the alignment of the chemical potential is done w.r.t. the one which fixes the value.

[395235863]

so in my second calculation(2nd pic),the chemical potential is the vaccum potential?needn't any shift?

[Anders Blom]

It's not so simple to just say "yes". You need to understand what the concept means, since it can be used for slightly different things. If this were a metal, then what you have computed is the work function = the distance from the vacuum level to the chemical potential. For a semiconductor, however, you can compute the electron affinity in the same way, or alternatively the distance from the valence band maximum to vacuum, but for that you also need to consider the band gap, since  the chemical potential in ATK is aligned mid-gap.