simulation of tunneling current in STM setup

[zhangguangping]

Dear all,

Can we simulate the tunneling current in a STM setup, where the tip is overhead the surface at a certain distance. However, when the distance of tip from the surface increases to 5 Angstroms or more, the tunneling current is zero no matter what the bias voltage is. However, in the experiments, the tunneling current is also observed when the tip is several nanometers away from the surface. Is this from the localized atomic orbitals?

[Anders Blom]

You could add ghost atoms in the vacuum gap to provide empty orbitals that the electrons can tunnel through.

[zhangguangping]

You could add ghost atoms in the vacuum gap to provide empty orbitals that the electrons can tunnel through.

Absolutely correct. That is what I am trying to do. Thanks for your reply. I saw the tunneling current when I add the floating orbitals. However, the paper (PRB, 79, 075441(2009)) says that "in the study of molecules adsorbed on surfaces, and especially for weakly interacting species, the use of floating orbitals can be problematic, since they could overlap with the molecular basis". Limited by my knowledge, what the problem will be if the floating orbitals are used for the molecule absorbing on the metal surface?  Similarly, if we want to get a more correct result, the atoms on the tip should also be surrounded by the floating orbitals (to get a better description of the wavefunction on  the STM tip) since it is also a kind of surface, yes?

[Anders Blom]

Without having the context, I can mostly guess, but I could imagine they are referring to the BSSE (basis set superposition error). This is mostly a problem when comparing total energies though, and I don't see that it would be an issue for the STM tunneling current.

The BSSE - and how to get around it - is discussed in the new tutorial http://quantumwise.com/documents/tutorials/latest/Grimme

[zhangguangping]

Without having the context, I can mostly guess, but I could imagine they are referring to the BSSE (basis set superposition error). This is mostly a problem when comparing total energies though, and I don't see that it would be an issue for the STM tunneling current.

The BSSE - and how to get around it - is discussed in the new tutorial http://quantumwise.com/documents/tutorials/latest/Grimme

Thank very much for your quick reply. According the tutorial, since an artificial attraction will arise as A and B get closer each system by utilizing the basis functions of the other system to lower its energy, we will get a smaller bond length (or distance) for A and B after optimization by using LCAO basis? The BSSE is used to correct the binding energy, so how to correct the optimized bond length error, if exists, arises from the LCAO basis orbital?

[Anders Blom]

Just to get the nomenclature right: the BSSE is not used to correct anything, "E" stands for error. BSSE is the error.

We use the counterpoise correction to correct for the BSSE, and this fixes both energies and geometries.

[zhangguangping]

Just to get the nomenclature right: the BSSE is not used to correct anything, "E" stands for error. BSSE is the error.

We use the counterpoise correction to correct for the BSSE, and this fixes both energies and geometries.

Thanks you. I see, I think it was a silly question.
Do you think  we can do the geometry relaxation of A in the present of B as ghost atoms? If yes, how to deal with ghost atoms? Fixed or relaxed either? That is, can the floating orbitals move during the relaxation?

Thanks very much for your attention.

[Anders Blom]

I doubt it will have any major effect unless the structure is far from equilibrium to begin with, but I would let them move, they will experience forces.

[zhangguangping]

I doubt it will have any major effect unless the structure is far from equilibrium to begin with, but I would let them move, they will experience forces.

Thank you very much for your reply.
Have a nice day.