what cause the saltations? Is the results reliable?

[EdonChen]

Dear all,

There is a question that puzzled me for a long time.

My work is about a porphyrin-like molecule attachs to Au(111) electrodes, which forms a Metal-Molecule-Metal junction.

See the attachments, the molecular orbitals shift versus applied bias and I-V curve, the saltations occur in I-V curve and orbital shift at bias  from ±1.3 V to ±1.35 V, specifically, from 1.32V to 1.33V at positive voltage, and -1.34V to -1.35V at negative voltage. It was strange that not all the orbitals saltated at the above appiled biases.

In the beginning, we thought it was miscalculation or the effect of exchange-correlation potential, but our re-calculation, and change of exchange-correlation potential still showed the  characteristic.

So, what cause the saltations? Is the results reliable?

Any reply for me is helpful, thank you.

[EdonChen]

Could anyone do me a favor?
It's really a urgent trouble for me.
Thank you again!

[kstokbro]

Dear EdonChen,
Looks like a very interesting effect you have found. Interestingly there seems to be only some of the orbitals are affected. I would investigate the difference in the orbitals affected and those not.  I would also look at the voltage drop in the device before and after the jump, my guess is that the drop shifts position and maybe the charges in the molecule shifts positions.

I noted that the effect seems to be related with the HOMO entering the bias window. In this case the occupation of the Homo might drop and instead the occupation of the LUMO and LUMO+1 could increase.

Thus, I believe it is a correct effect of the calculation, hope my suggestions for investigating it helps.

[EdonChen]

Thank you very much, professor Kstokbro, I am sincerely appreciate your help. And your  affirmation about the results makes me so excited.

Exactly, I've tried to find out the differences of the affected orbitals(LUMO&HOMO-2) and those not(HOMO&LUMO+2). But the MPSHs didnot show many remarkable differences, only HOMO-2 showed a difference of localized degree. And I also have tried to look into the differences of LDOS, PDOS, Effective Potential and electrostatic difference potential, Mulliken Population etc. before and after the jumps, but seems all fail to explain such phenomenon.
And I followed by your advice to calculate the voltage drop, as seen in attachment, since I am not quite familiar with voltage drop, Dear Kstokbro, do you think whether it might tell something?

"I noted that the effect seems to be related with the HOMO entering the bias window. In this case the occupation of the Homo might drop and instead the occupation of the LUMO and LUMO+1 could increase."
I don't quite understand this, because the HOMO has already entered into bias window before the jumps occur, in my GGA calculation, this jumps shift to ±1.55 V to ±1.60 V, while the HOMO still entered into bias window at ~±1.2V. So, I am anticipating your profound interpretation.

Thank you again for you kindly help!