About contact resistance

[hasanahmad]

I used (7,0) CNT with gold electrodes to see the contact resistance. I found out that using different repetitions gave me different result. Can you please tell me why this is happening as the CNT is ballistic and I just changed the number of unit cells (UC) keeping the distance at ends, the same.

I used 5 UC and got ~45k, but when I used 2 UC, I got ~18.2k. Kindly need your help in this regard.

[mads.engelund]

Could you please provide a script?

[hasanahmad]

I used builder to setup the geometry. The number of repetitions used were 2 and 5 for the different cases. I used DFT with the default settings for self-consistency and for transmission calculation.

[hasanahmad]

Sir,
I tried another length to see whether it corresponds to any of the contact resistance calculated. I used 20 repetitions of (7,0) CNT with same lengths from Au contact and got 3.1M ohm as the resistance.

Kindly can you please tell me what I am doing wrong
The three device structures can be seen in the figure below.

[nori]

I think your calculation is reasonable.
Because (7,0)CNT is a semiconductor, there is no (quasi) propagating mode of (7,0)CNT within a bias window at least when a bias voltage is low, so that electron transport occurs via only decaying modes.
In other words, all scattering states within a bias window are decaying modes of (7,0)CNT coupled with propagating modes of Au electrods.

[Anders Blom]

If you really want to model the contact resistance you should instead set up the system as a single interface, just metal|CNT. In your case you have first of all two contacts, metal|CNT|metal, and as Nori correctly pointed out (7,0) is semiconducting so for sure there will be an exponentially decaying current. So you are not computing the contact resistance but the conductance of a thin metal-semicondcutor-metal junction, and in this case it's perfectly expected that the conductance depends on the length of the semiconductor part.