the question about the gate voltage in ATK

[jiguomin]

Hello,  everyone
I study the effect of gate voltage on the electronic transpoort of molecular device. The gate voltage is applied on the left of the molecule only. The reviewer ask the below queation:
It is unclear from the paper how the 'gate' voltages translate into orbital energies.  Did the
author’s take the voltage of  'gate' and apply that uniformly to the atoms of the 
molecule?  If this is the case then the result is not physical, since each gate will contribute to
the energy shift on all atoms nearly identically (since the capacitance between each gate and
atom is nearly identical).  This is a major problem with this manuscript.
How can I answer? Thanks.

[mads.engelund]

The gate voltages are not directly applied to the orbitals. That would indeed be unrealistic. The gate voltage is applied to a region of space away from the molecule and the electrostatic equation is solved as an integrated part of the NEGF-DFT(Slater-Koster and Huckel too) method.

Take a look at this article.
http://prb.aps.org/abstract/PRB/v65/i16/e165401

[mads.engelund]

This also means that the position and geometry of the gate is just as important as the gate voltage. The closer the gate is to the atoms, the bigger effect it has.

[Anders Blom]

The above is correct, but it also depends on how the "gate" was applied. An alternative way to apply it (indicated perhaps by your notion that it's only applied to part of the system) is to introduce a shift of the onsite energies (the old "gated method" in ATK 2008.10 and earlier, and also still possible in 12.2 and later, but with another method). In this latter case I would tend to agree with the reviewer that it's unrealistic to assume that some atoms in a very small region are "gated" and other ones not - if, that is, you really mean this to simulate an external gate. There is also internal gate effects, more properly I guess charging effects which can be caused by reduction/oxidation of individual atoms or side groups. This charging effect can be simulated using an onsite shift to some extent (and in the upcoming ATK 13.8 but a real localized charge on an atom). Ultimately, you have to be able to argue your case why you have simulated the system in a particular way and what physical relevance it has, otherwise the study has no meaning.

[jiguomin]

The above is correct, but it also depends on how the "gate" was applied. An alternative way to apply it (indicated perhaps by your notion that it's only applied to part of the system) is to introduce a shift of the onsite energies (the old "gated method" in ATK 2008.10 and earlier, and also still possible in 12.2 and later, but with another method). In this latter case I would tend to agree with the reviewer that it's unrealistic to assume that some atoms in a very small region are "gated" and other ones not - if, that is, you really mean this to simulate an external gate. There is also internal gate effects, more properly I guess charging effects which can be caused by reduction/oxidation of individual atoms or side groups. This charging effect can be simulated using an onsite shift to some extent (and in the upcoming ATK 13.8 but a real localized charge on an atom). Ultimately, you have to be able to argue your case why you have simulated the system in a particular way and what physical relevance it has, otherwise the study has no meaning.

Thanks for your reply. I use the ATK 2008.10. I apply the gate voltage on the first and second molecule of the scattering region, respectively. I want to investigate the effect of gate voltage applied the different positions on the electronic transport properties of molecular device. How to answer the question raised by the reviewer? Thanks a lot.

[jiguomin]

This also means that the position and geometry of the gate is just as important as the gate voltage. The closer the gate is to the atoms, the bigger effect it has.

Thanks a lot for your reply. I use the ATK 2008.10. I applied the gate voltage on the first molecule and the second molecule of the scattering region, respectively. I just study the effect of gate voltage applied different positions on the electronic propersities of the molecular device. I want to know how the 'gate' voltages translate into orbital energies.  Did the ATK 2008.10 apply the gate voltage uniformly to the atoms of the molecule?  If this is the case then the result is not physical, since each gate will contribute to the energy shift on all atoms nearly identically (since the capacitance between each gate and
atom is nearly identical).

[Anders Blom]

The "gate voltage" was included as an onsite diagonal term, so yes, all atoms (or rather, all atoms selected to be "gated") are affected equally. The models in later versions of ATK are more physical since you can include actual physical gates which influence the atoms through the electrostatic field.

[jiguomin]

This also means that the position and geometry of the gate is just as important as the gate voltage. The closer the gate is to the atoms, the bigger effect it has.

Thanks for your reply. I use the ATK 2008.10. There is only the paramater that the atom number applied of voltage gate. How to understand "The closer the gate is to the atoms, the bigger effect it has"

[Anders Blom]

In 2008.10 this is not relevant, since all atoms will be affected equally by the "gate". As mentioned, in newer versions of ATK the model of a gate is more realistic.