question about the zero energy point in the transmission spectrum

[beark]

Hello, everyone! I have met a problem about the position of zero energy point  in the transmission spectrum.
I tested a 1-D nanoribbon device. The nanoribbon is devided into two electrode regions and a scattering region.
Thus I think the transmission spectrum under zero bias can be totally determined by the band structure of the
nanoribbon. But the result turned out not to be what I thought.... (see figure 1)
However the shape of the transmission curve seems reasonable if I shift it upward. (see figure 2)
I guess the zero energy point in the transmission curve is a relative value.
But what's the physical meaning of the zero energy point in the transmission curve?
Is it means the fermi energy of the device ? And what's the difference between the fermi energy from the band
structure and the two probe device?





 

[zh]

The zero of energy in the transmission spectrum is set as the Fermi level which is defined as the average value of the Fermi levels of two electrodes.  The transmission spectrum is not the DOS in which the peak may correspond to dense or flat energy bands.

The detailed discussion about the Fermi level of two-probe system can be found in this thread:
http://quantumwise.com/forum/index.php?topic=120.0

[beark]

Thank you, zh! I have read the post. I think you mean the relative electrostatic potential  can be shifted by
a random constant number and it's different for the bulk system and two-probe system?
But the system I calculated is homogeneous, and it's under zero bias.
And I have calculated other homogeneous system by using ATK before. The band and the transmission curve
match well....

[zh]

The possible reason may be that the Fermi levels of electrodes are not determined properly. The temperature parameter in the Fermi-Dirac function and the size of k-grid is required to be checked carefully.

[beark]

Thank U, I will check them carefully.
I found that although the transmission spectrum seems strange, the current
seems not influenced by the unphysical shift of transmission curve. The I-V curves
go on well under finite bias.
Thus I guess the fermi energy determined in the current calculation is different from
that in the transmission calculation? 

[Anders Blom]

Is this computed using ATK-DFT or ATK-SE?

Generally, there is no difference in the zero-point definition, but in ATK-SE at least there is a possibility to specify separately the zero-point in the band structure and the transmission, so if they are not set to the same, then you can see this effect. Perhaps you can post your script, so and we can have a look.

[beark]

It is using ATK-DFT.
The parameter part of the script which may have some problems is posted.

[Anders Blom]

There is not enough information here to judge what the issue might be. Basically, what is unclear is how the structure looks (because the band structure you show is unusual for a simple nanoribbon), and how the band structure is plotted.

If you prefer not to share your geometry publicly, you can send it privately to me.