Dear Mohit, The green's function as implemented in kwant returns the elements of this latter only between the sites connected to the leads. The idea behind mounting fictitious leads with zero self-energy on some specific sites inside the system is just to tell kwant that I want the Green's function for these sites too. Now, in the answer I gave to you last time, I supposed that your system is finite and had no leads (other than the fictitious leads you mounted). That is why I said: the current will be zero if the fictitious self energy is zero. Is your system finite? I also do not see from where did you get this formula.((H_ij G_ij-H_ji G_ji)) ? The usual formula for the current is Im[H_ij (G Gamma G^\dagger)]_ij, Gamma = 2Im[Self_energy]. So please check also your formula. Most importantly, Kwant has functions which return local current which are very efficient. Why do you not use them? Now, I come back to your question about the meaning of what the function of greensfunction returns: If the mounted lead is number 3, and you want the greens function between sites site1, site2 than you do: mount_vlead(syst,[site1,site2],2) G=greens_function(syst.finalized(),energy=-1.8, in_leads=[3],out_leads=[3]).data This will return the matrix G=([G11,G12],[G21,G22]) if the number of orbitals =1 otherwise it will double in case norbs=2. G12 id the greens function element from site 1 to site 2. I hope this helps. Adel On Wed, Jun 17, 2020 at 3:02 AM <mohit.gupta9607@gmail.com> wrote:
Dear Abbout, Thanks for your reply. The idea was to evaluate green's function for two particular slices of the system I basically followed this discussion: https://rb.gy/v6bkpo Which says to add zero self energy to obtain green's function everywhere. So my next question is if I don't give zero self energy won't my Green's function depend on the choice of Gamma?
I think the other issue is that I am not 100% sure about how to read the output of the greens_function(syst,energy).
-- Abbout Adel