QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: donmbringer on September 19, 2012, 13:29

Title: About the molecular wires junction.
Post by: donmbringer on September 19, 2012, 13:29

Hi,

I built the molecule junctions with a benzene completely following the tutorial and then calculated the I-V curve. May I ask, whether this calculation includes the tunneling (or noise) current between two electodes, because the space between two electrodes are very small?

Thanks ahead.

Title: Re: About the molecular wires junction.
Post by: Anders Blom on September 19, 2012, 13:39

No, there is no direct tunneling between the electrodes, if there is then the model breaks down. So all electrons have to go through the molecule.

Title: Re: About the molecular wires junction.
Post by: donmbringer on September 19, 2012, 13:42

Thanks for your kind reply.

I have one more question: it seems that there is still considerable difference between the literature (1.jpg)  (Reed et al. Science, 1997) and the simulation. The simulated results (2.jpg) are around 30 times bigger than the literature. The shape also has a little bit difference between them. Could you please give me some suggestion on it?

Title: Re: About the molecular wires junction.
Post by: Anders Blom on September 19, 2012, 22:39

I can't really comment at all, since I know nothing of your calculations, parameters, geometry, etc. It's hard in general to demand a perfect quantitative match because there are so many factors in the experiment that you don't include in a model like this, not least a "simple" thing like the Au-S distance which influences the conductance exponentially.

Oh, you mean your calculation was exactly following the tutorial, sorry, I missed that. Well - the second part of the comment still holds :)

Title: Re: About the molecular wires junction.
Post by: Anders Blom on September 19, 2012, 22:43

Oh, you mean your calculation was exactly following the tutorial, sorry, I missed that. Well - the second part of the comment still holds :)

To reiterate the point in a different way, what you should look for in the simulation results are the qualitative features that match the experiment, such as the peak and then decrease in conductance at 1.2 V, which is reasonably well reproduce, perhaps not so good for higher bias, but still the feature is present in the calculation too. In the experiment it can be very hard to explain such effects, but what the simulation can contribute is a proper understanding of the phenomenon.