Re: [Kwant] transfer-matrix method
Dear Michael, Thank you for the prompt reply! Yes, I mean the possible edge modes if I cut the lead. The edge modes depend on the "vacuum" that is outside the lead. One trick to get some edge modes (suggested to me by Gene Mele) is to find parameters when right-decaying and left-decaying evanescent modes become linearly dependent. "glue the lead to it's mirror image". There may be other ways, dependent on the boundary conditions.
Best wishes, Sergey
On 29/07/15 12:04, Michael Wimmer wrote:
we haven't at the moment published a paper with all the details of the algorithm used in kwant. I discussed some very much related methods in my thesis (then in terms of Green's functions, in kwant we extended this to wave functions) which you can find at
How to find the propagating and evanescent modes is described there (in kwant we have a smarter way of doing it for non-invertible hopping matrices, but I believe this is not too important for you).
I'm not entirely sure what you mean with edge states? Do you mean states sitting at the end of the lead, if I decouple the lead from the scattering region? Those appear whenever the evanescent modes become linearly dependent. Is that what you are asking for?
On 29-07-15 12:39, Sergey wrote:
Dear Developers, Could you give a link to the paper where the transfer-matrix method used in Kwant to find the propagating and evanescent modes in the lead is discussed? Can this method produce edge states in the lead? (can these related to evanescent modes?) Best wishes, Sergey