Dear Gabriele, It's very likely as you said. I took a (12,0) 5 layers long for the system and a 1 layer long for leads; I expected not have neglegible effect for a bias of 0.4 V and be able to cunclude if running transiesta is necessary or not. I should admit that I thought this comparison would be good since I found resonable the electronic properties are different for different bias. In this particular test-system, I think everything must happen in the contacts since it's a perfect system. I will do some checks in order to be sure and perform the transiesta calculations for every bias as you suggested for the more "interesting" cases. It was very enriching this discussion with you and Guanping.
Kindest regards Juan Manuel On Fri, Apr 13, 2012 at 9:12 AM, Gabriele Penazzi <g.pena...@gmail.com> wrote: > Dear Juan Manuel, > > Guanping explanation is right. I would say in general, it should be the > same whether the local charges in the system don't change significantly > under bias. > This is more likely to happen in a system like a CNT than, for example, > in a single molecule device, where you should start to see relevant > different already for a small bias. > So there's not really a yes or no answer, it actually depends on your > system > > Regards, > Gabriele Penazzi > > > > On 04/12/2012 04:52 PM, Juan Manuel Aguiar wrote: >> Dear Siesta Users, >> I have a very simple question for those regular users of transiesta >> and tbtrans: for plotting IV characteristics; do I have to run >> transiesta for every bias? >> I've done a test with a small system (a single wall CNT) and I found >> no difference between running transiesta and then tbtrans for every >> bias or just running transiesta for 0 bias and then tbtrans for all >> the bias values. >> I think that this question is answerd easyly: yes or no. Let me know >> if I had to deliver details of the calculation I expect to do. >> >> Juan Manuel > >
