QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: Sarvesh Agarwal on October 27, 2012, 08:39

Title: metal_nanotube
Post by: Sarvesh Agarwal on October 27, 2012, 08:39

hello !!!!
          in order to calculate the IV characteristics for a metal_CNT(5nm)_metal using extended huckel, when i find the transmission spectrum b/w (-2,2)ev it shows negative  current for positive voltage.
          then to correct that i increased the range for the transmission spectrum from (-8,16)eV now its giving positive currents... so please tell me is it correct??

Title: Re: metal_nanotube
Post by: Anders Blom on October 27, 2012, 11:29

It's not really enough to just look at the current, you should inspect the transmission spectrum itself. If this is negative (which is unphysical), then you can get this effect, which existed in older versions of ATK (11.8 and before). To work around it, try using the Recursion method instead of Krylov when computing the transmission spectrum. It's slower, but numerically more stable.

Title: Re: metal_nanotube
Post by: Sarvesh Agarwal on October 28, 2012, 16:43

thanks sir!!
   please tell me getting negative transmission spectrum makes any sense.. or it is totally non-physical and it only means that my script is somewhere wrong..

Title: Re: metal_nanotube
Post by: Anders Blom on October 28, 2012, 21:12

It is unphysical, but probably not due to any error in your script (although it's possible), but rather due to the Krylov method used to calculate the transmission spectrum. Sometime it fails numerically, which can give negative transmission. In 12.2 and later we have fixed this (in as many cases as possible), but as I wrote above, if you want to be on the safe size, use Recursion or Direct instead. It's slower, however.

Title: Re: metal_nanotube
Post by: Sarvesh Agarwal on October 29, 2012, 07:02

thanks!!! sir