total Charge is not conserved in two probe SCF calculation

[jjhskang]

Hi!

I am making two probe calculations using atk2008.10.

In one of them, I find that the SCF calculation at zero bias gives
 
 q =  412.42189 e  Etot = -2064.51856  ...... at the end of the convergence iteration.

However, the total # of electrons in the scattering region should be 413, and some electrons disappeared.
What is the problem here and how can I solve this problem?

[jjhskang]

(1) Is it because I have not included all the poles in the calculation?
I used the default value of  integral_lower_bound, i.e., 3eV.
 
(2) I believe that I have inlcuded farely large number of atoms in the scattering region in order to ensure completel screening in the scattering.

Hi!

I am making two probe calculations using atk2008.10.

In one of them, I find that the SCF calculation at zero bias gives
 
 q =  412.42189 e  Etot = -2064.51856  ...... at the end of the convergence iteration.

However, the total # of electrons in the scattering region should be 413, and some electrons disappeared.
What is the problem here and how can I solve this problem?

[Anders Blom]

The two-probe calculation in ATK is carried out under open boundary conditions. That means, that charge is not conserved in the central region, and this allows for finite-bias non-equilibrium calculations.

[kaypu]

the lost of the charge can be considered as the alignment of the Fermi energy,
alignment means the Fermi energy of the electrode and the scattering area, is that right?

[Anders Blom]

Basically, yes.

Do note however that since the electron distribution in the central region by definition is not in equilibrium, there is not, strictly speaking, any well-defined Fermi level in this region. It's position-dependent, to lowest order perhaps similar to a linear ramp between the applied biases, but formally it cannot be defined.