spin

[dwh]

Hello everyone
 I calculate the bandstructure of graphene nanoribbon with considering spin,but my result is different from others that had published. There are the relevant plots and py file below. I do not know that my py file si wrong or right when i consider the spin for the system,could you test my py file ?

[zh]

In your script file, you did a spin-polarized calculation for the ferromagnetic state of zigzag graphene nanoribbon.  In addition,  4 x 1 supercell of zigzag graphene nanoribbon  is used. So, the band dispersion of two bands near the Fermi levels are folded in your calculation results. This makes your results different from the published one, which is obtained by the primitive unit cell of zigzag graphene nanoribbon. 

Please do the anti-ferromagnetic state of zigzag graphene nanoribbon using its primitive unit cell. You will  reproduce the published result exactly.

[dwh]

Thanks for your reply,but I reproduce it again and the rusult is still different from others that had published(N.K. Jaiswal, P. Srivastava / Solid State Communications 152 (2012) 1489–1492). The bandstructure of 4 x 1 supercell of zigzag graphene nanoribbon is calculated in the published ,not a primitive unit cell. There are some the relevant information in the plots below about the published paper.
I calculate the bandstructure with 4 x 1 supercell of zigzag graphene nanoribbon and a primitive unit cell respectively. And the result is at additional options.
I am looking forward to your reply again,thanks very much.

[dwh]

This is the other informations about the published paper and my script file

[zh]

For the band structure of the pristine case, the authors definitely used the primitive unit cell rather than the 4 x 1 supercell. For doped case, the authors may used the 4 x 1 supercell. Actually, you results obtained by the primitive unit cell is the same with the band structure of pristine case in the published paper.

[dwh]

Thanks for your reply. I will continue to reproduce that paper result and I come to another question that I calculate the DOS (including PDOS) of zigzag graphene nanoribbon primitive unit cell .The result below shows that the curve is very sharp rather than smooth.How can I do in order to get the smooth DOS curve?

[Nordland]

First of all you should use much more k-point, but it is important to use them properly - you are using 10x10x10 in your script ( and in the earlier post ) you had 100x100x100, but this is just a waste of computation power in your systems as it is a 1d system. You should use 1x1xN and you could safely use N=1000 or N=5000, and you curve should be much more smooth.

[zh]

To get smooth DOS curve, you may proceed the broadening by means of the Gaussian or Lorentzian functions. The 1 x 1 x N (N <= 100) may be sufficient for the k-point sampling in the DOS calculations of zigzag graphene nanoribbon.

[dwh]

Thanks,but i do not know how to use the the broadening by means of the Gaussian or Lorentzian functions.Could you give me more detail information or example of script file. thanks

[Anders Blom]

For bulk systems, ATK uses the tetrahedron method which doesn't really use or need broadening, as long as you have enough k-points. Follow Nordland's advice above and you should be fine. It is however possible to evaluate the DOS with a Gaussian broadening, if you really want, see the reference manual for details.