Mailman 3 December 2024 - Kwant-discuss

Abort trap: 6 error message
by Marc Vila Tusell 18 Dec '24

18 Dec '24

Hi, I recently got a new computer, MacBook Pro, Sonoma 14.6.1 with chip Apple M3 Pro, and I installed kwant with Conda, as I did in the past many times in other computers (both Mac and Ubuntu). The installation seemed to had succeeded. However, when I do "import kwant", an error appears. Specifically when I do it in the python prompt of the terminal, the python prompt exits and then the following line appears "Abort trap: 6". I have tried to search for this error in relation to python packages but I haven't been able to find something that can help me. Additionally, by trying to understand and fix this, I found the following. If I go the the solvers folder where kwant is installed within the Conda environment, and I do "import kwant" again in the python prompt, kwant gets imported but I get the following message: /Users/XXX/anaconda3/envs/kwant/lib/python3.12/site-packages/kwant/solvers/default.py:16: RuntimeWarning: MUMPS is not available, SciPy built-in solver will be used as a fallback. Performance can be very poor in this case. warnings.warn("MUMPS is not available, " I do have mumps installed (mumps-include, mumps-seq, python-mumps) when I check with "conda list mumps", so I don't really understand what is going on. I figured the most possible explanation may be related to the fact this Mac has an apple silicon chip, as this sometimes gave me problems in some other situations. However, I have not found anybody in this mail list posting error messages related to apple silicon chips. Any help and direction will be greatly appreciated. Best, Marc

1 0

particle-hole symmetry in the tutorial sec2.6
by wilson2048＠outlook.com 15 Dec '24

15 Dec '24

Dear kwant users: I am reading the tutorial `sec2.6` and encounter the disscussion about the particle symmetry. (Note you can put the following text into the editor which supports markdown for better format, like jupyter.) From the given code, it seems that we have $S_{ee}=S_{hh}^*$ and $S_{eh}=-S_{he}^*$: ``` def check_PHS(syst): # Scattering matrix s = kwant.smatrix(syst, energy=0) # Electron to electron block s_ee = s.submatrix((0,0), (0,0)) # Hole to hole block s_hh = s.submatrix((0,1), (0,1)) print('s_ee: \n', np.round(s_ee, 3)) print('s_hh: \n', np.round(s_hh[::-1, ::-1], 3)) print('s_ee - s_hh^*: \n', np.round(s_ee - s_hh[::-1, ::-1].conj(), 3), '\n') # Electron to hole block s_he = s.submatrix((0,1), (0,0)) # Hole to electron block s_eh = s.submatrix((0,0), (0,1)) print('s_he: \n', np.round(s_he, 3)) print('s_eh: \n', np.round(s_eh[::-1, ::-1], 3)) print('s_he + s_eh^*: \n', np.round(s_he + s_eh[::-1, ::-1].conj(), 3)) ``` However, from my naive reasoning I would expect that $S_{ee}=S_{hh}$ and $S_{he}=-S_{eh}$. Here is my reasoning, $$ H \begin{pmatrix}\psi_e \\ \psi_h \end{pmatrix}=E\begin{pmatrix}\psi_e \\ \psi_h \end{pmatrix} \Rightarrow \begin{pmatrix}\psi_e^{(out)} \\ \psi_h^{(out)} \end{pmatrix} = S(E) \begin{pmatrix}\psi_e^{(in)} \\ \psi_h^{(in)} \end{pmatrix} $$ With partilce symmetry $\sigma_y H \sigma_y = -H$, we have: $$ H \sigma_y \begin{pmatrix}\psi_e \\ \psi_h \end{pmatrix}=-E \sigma_y \begin{pmatrix}\psi_e \\ \psi_h \end{pmatrix} \Rightarrow \sigma_y\begin{pmatrix}\psi_e^{(out)} \\ \psi_h^{(out)} \end{pmatrix} = S(-E) \sigma\begin{pmatrix}\psi_e^{(in)} \\ \psi_h^{(in)} \end{pmatrix} $$ which leads to $$ S(E) = \sigma_y S(-E) \sigma_y, $$ giving $S_{ee}(E)=S_{hh}(-E)$ and $S_{he}(E)=-S_{eh}(-E)$. Best regards, Wilson

1 0

When will be the next update of tKwant?
by Zeng chuanchang 14 Dec '24

14 Dec '24

Hi, I am trying to use tKwant to do some numerical simulation, and then I found it is not completely self-consistent/ done currently, as some of the given codes/examples can not be complied with and run successfully. So, I am wondering when the next update will be. Looking forward to hearing from you soon! Best wishes, Chuanchang Zeng zengcc(a)baqis.ac.cn Beijing Academy of Quantum Information Sciences Beijing, China

4 4

University profs in US using kwant etc
by Konner Feldman 14 Dec '24

14 Dec '24

I'm an undergraduate studying physics at UT Austin and looking for graduate schools. For the past 18 months or so I’ve been working with Kwant on the Frontera supercomputer using it to simulate Josephson Junctions. I’m interested in continuing working in computational physics/ condensed matter. Does anyone know of professors at Northwestern, Yale, Princeton, MIT, NYU, University of Illinois or similar schools that are using Kwant or something similar for computational physics research? Thank you!! Konner

2 1

conservation_law and particle_hole in finite systems
by Konner Feldman 14 Dec '24

14 Dec '24

I am trying to simulate a finite size rectangular 2d electron gas inside a josephson junction. There is one rectangular scattering region on each end that is coupled to a different superconductor, while the middle rectangular scattering region is just a part of the semiconductor 2DEG. I am only simulating the 2DEG and not the superconductors, but there is still spin-orbit coupling and hole interactions. I was told that the "conservation_law" and "particle_hole” parameters are not needed for systems such as this one that do not have infinite length regions, but I don't understand why. Could somebody please help me understand this?

2 2