QuantumATK Forum
QuantumATK => General Questions and Answers => Topic started by: bjnagare on December 19, 2009, 15:02
-
My system of two probe (Graphene-Graphene-Graphene) is converging well with (1,1,50) kpoints with msrh cutoff of 150Ry and single zeta polarized. The same system is not converging with (10,1,100) kpoints. Kindly suggest some guidance.
Thank you
-
You may try to increase gradually the size of k-point mesh and do a restart calculation from a previous converged scf calculation with smaller k-point mesh.
-
Is it pure 2D graphene in the electrodes? In that case you will want to take care to include the K point in the k-point sampling. Try 9 k-points instead, or some other multiple of 3 (although it's not always the K point is included anyway).
-
Dear Sir,
Yes. It is pure 2D Graphene.
But I haven't understood the concept of K point sampling with multiple of 3.
Thank you
Nagare
-
The K point mentioned here is ".333333333333333, .333333333333333, .000000000000000" in lattice coordinate format (http://dave.nrl.navy.mil/bind/kpts/format.html). This is due to the peculiar band structure of graphene, i.e., the 2 bands near the Fermi level due to the [tex]\pi[/tex] and [tex]\pi^{*}[/tex] states are crossing around the K point. For example, pls see here:
http://www.iue.tuwien.ac.at/phd/pourfath/node18.html (http://www.iue.tuwien.ac.at/phd/pourfath/node18.html)
-
Dear Sir,
Thank you very much for your helpful suggestion.
Regards
Nagare
-
Dear
It seems that you are not satisfied with the pure 2D graphene in the electrodes. What is your seggestion about setting electrodes for graphene if I want to do this job in ATK 2008 ?
In ATK-SE manule, eamples about big system is little, hopping to see more .