caculation of bloch state

[mldavidhuang]

When I calculate the bloch state of zigzag graphene nanoribbon using LDA, I found there is no item of 'none' for spin panel as in the tutorials. what is the problam?
And after I calculated the one spin bloch state of the lowest conduction band at Z, I found the one spin also localized on the edge of ribbon, just as the result using LSDA with the inital setting of edge spin polarizations. Is that right?

[Anders Blom]

You may have an older version of the tutorial, or we might have forgotten to update a screen shot. There shouldn't be an option None, choose Up if the calculation is not polarized.

[mldavidhuang]

But I got the result that the Z point's bloch state localized on the edge of ribbon when the calculation is not polarized. Is it reasonable?

[zh]

For the zigzag graphene nanoribbons, the edge states can be produced either by the spin-polarized calculation or by the non-spin-polarized one. Your results are reasonable.

[Anders Blom]

Yes, the only difference is that with spin the two spin components separate on opposite edges.

[mldavidhuang]

But the result I got from the non-spin-polarized one is that the edge states are localized on one edge of the ribbon for lowest conduction band while , for highest valence band, on other edge.

[Anders Blom]

Are you sure it's not polarized? If you set initial spins, but used LDA (not LSDA), it's still polarized.

[mldavidhuang]

Yes, I am sure... Maybe you can try with new version with unit cell of zGNR, it is quite fast

[zh]

Just now I did a test calculation for a zigzag graphene nanoribbon. The results obtained by from the LDA calculations are different from what you have stated, i.e., the Bloch state for the highest occupied eigenvalue of Z point (0, 0, 0.5) is localized at both edges of ribbon.

[mldavidhuang]

I did it once again, and found the result is the same as yours. I am wrong, sorry for making such mistake!