QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: Arya on February 18, 2014, 03:34

Title: Skipping variation of potential in X, Y direction
Post by: Arya on February 18, 2014, 03:34

Hi,

Greetings !

In the NEGF formalism (what I understood by reading some papers from S. Datta) we use a eigenspace representation for Hamiltonian in X, Y direction and a real space Hamiltonian representation in transport direction.
This means we need the variation of potential (U) only in transport direction since XY direction are taken care by eigenspace representation.

If my understanding is correct do we have option of not calculating potential variation in XY direction, thus speeding up the calculations.

Of course others might want to see a 3D view of potential but in case one doesn't need; can we skip the X,Y direction and use just the Z direction for NEGF calculations.

Thanks,
-Arya

Title: Re: Skipping variation of potential in X, Y direction
Post by: kstokbro on February 18, 2014, 07:20

you can use periodic boundary conditions in the x,y directions, in this way you can have a very small cell in these directions and significantly speed up the calculation. However, you need to include k-point sampling then in x, y, since even if you potential is periodic the wave functions are not, they are Bloch waves, check Blochs theorem.