Dear Andreas,

I investigated your code and everything seems correct. I checked your claim, and found it true! For large value kapp=0.0001, the S matrix is unitary. The same for Kappa=0 or kappa <1E-13. Around Kappa=1E-9, it is not unitary (at least not unitary with the precision set in kwant) I am still puzzled and did not find an answer for that. The only remark I can make is that the edge state at this energy is degenerate and I remember, a case with kwant having a problem happening at the degenerate states.

If it happens that you find the answer, I will be interested to read it.

Best regards, Adel

On Thu, Dec 10, 2020 at 5:49 PM Andreas Bereczuk Bereczuk.A@gmx.de wrote:

Dear all,

i used a slightly modified "discretize.py" code of https://kwant-project.org/doc/1/tutorial/discretize (see attached file)

and added the term

""" + kappa * kron(sigma_z,sigma_0) """

to the hamiltonian (splitting up the different spins). When checking the unitarity of the scattering matrix at energy=0 by applying the function "check_unitarity(matrix)" for

kappa=0 and kappa=10**(-9) the unitarity-error is increasing from order 10**(-13) to 10**(-6) !

Why is the error of $ S^\dagger S$ getting so high?

Thanks in advance!

Andreas Bereczuk