Mailman 3 May 2015 - Kwant-discuss

Greens function
by 李航 11 May '15

11 May '15

Dear Authors, Thank you very much for supplying a so powerful tool to study the quantum systems. A series of complex physical phenomena can be modeled by Kwant. I would like use the Kwant to calculate the charge density in a quantum wire. I met some trouble in understanding the output of retarded Green function when I use the module with the name of “kwant.solvers. default.green_function.” Assume we consider an ideal quantum wire, which includes two leads and a scattering region. The width of scattering regime is identical to the leads. The scattering region has 4*3 discrete points. The total green function should be a 12*12 matrix and contain 144 elements. However, I only got 6 by 6 matrix and 36 elements. What am I missing? My code was the same as the code of the first example for quantum wire in page 16 in kwant document but with L=4 and W=3, energy = 1 and call for the module. greenf=kwant.solvers.default.greens_function(sys, energy) print greenf Outputs are: GreensFunction(data=array([[-0.28713703-0.20513305j, 0.08984532-0.29010194j, 0.16007656-0.20513305j, 0.26900184+0.11162436j, 0.20092781+0.07893034j, 0.17600545+0.07893034j], [ 0.08984532-0.29010194j, -0.12706047-0.4102661j , 0.08984532-0.29010194j, 0.37693326+0.15786068j, 0.26900184+0.11162436j, 0.26900184+0.11162436j], [ 0.16007656-0.20513305j, 0.08984532-0.29010194j, -0.28713703-0.20513305j, 0.26900184+0.11162436j, 0.17600545+0.07893034j, 0.20092781+0.07893034j], [ 0.26900184+0.11162436j, 0.37693326+0.15786068j, 0.26900184+0.11162436j, -0.12706047-0.4102661j , 0.08984532-0.29010194j, 0.08984532-0.29010194j], [ 0.20092781+0.07893034j, 0.26900184+0.11162436j, 0.17600545+0.07893034j, 0.08984532-0.29010194j, -0.28713703-0.20513305j, 0.16007656-0.20513305j], [ 0.17600545+0.07893034j, 0.26900184+0.11162436j, 0.20092781+0.07893034j, 0.08984532-0.29010194j, 0.16007656-0.20513305j, -0.28713703-0.20513305j]]), ************************************************************************************************************************** ************************************************************************************************************************** lead_info=[array([[-0.44909123-0.15234015j, -0.19564002-0.21544151j, -0.06712522-0.15234015j], [-0.19564002-0.21544151j, -0.51621644-0.3046803j , -0.19564002-0.21544151j], [-0.06712522-0.15234015j, -0.19564002-0.21544151j, -0.44909123-0.15234015j]]), array([[-0.51621644-0.3046803j , -0.19564002-0.21544151j, -0.19564002-0.21544151j], [-0.19564002-0.21544151j, -0.44909123-0.15234015j, -0.06712522-0.15234015j], [-0.19564002-0.21544151j, -0.06712522-0.15234015j, -0.44909123-0.15234015j]])], out_leads=[0, 1], in_leads=[0, 1]) Regards, Hang

2 1

Construction of scattering matrix
by Lee Kwok-Long 10 May '15

10 May '15

Dear all, I want to construct the scattering matrix for each propagating mode like: for kx=2.9, S=[(r,t)T，(r’,t’)T] , here all the matrix elements are numbers not a matrix. I think t_M can be expressed as t_M=smatrix.submatrix(1,0), which concludes all the propagating modes, similarly, r_M, r’_M and t’_M can all be written down. However, the problem is: how to find the r,t,r’ and t’ for the same propagating mode. For example, do t_M(1,1) and r_M(1,1) correspond to the same propagating mode? Is there a method to solve my problem? Thanks in advance! Best wishes， Kwok-Long Lee

1 0

Re: [Kwant] [PROVENANCE INTERNET] Re: bug in kwant installation requirement in PyPi
by Joseph Weston 04 May '15

04 May '15

Hi, > Thanks for letting us know about the problems with “pip”. > The state of python packaging has been messy, only recently there has > been some progress [1]. The way Kwant is packaged reflects the > documentation on Python packaging for Python 2.7. We’ll see how we can > improve things. I think the fundamental problem is that kwant doesn't declare it's dependencies in the standard way to setuptools using `install_requires`: https://packaging.python.org/en/latest/requirements.html Unfortunately it's not a 1-line fix just to add the dependencies, as numpy is imported within the setup.py script itself (so that we can get the numpy C headers; used for compiling the kwant modules written in Cython). I can't see a way around this using the current tools. Joe

2 1

bug in kwant installation requirement in PyPi
by Rafał Skolasiński 04 May '15

04 May '15

Hi All, I recently noticed that kwant installation available on PyPi via pip lacks requirements. In completely clean and new virtualenv the following fails: ''pip install kwant'' To make it work one need to install first ''pip install numpy cython'' and to run kwant: ''pip install scipy tinyarray'' Best, Rafal

2 1

attaching 1D lead to 2D lattice
by Sergey 02 May '15

02 May '15

Dear Colleagues, I want to measure the decay of wave-function in magnetic field. For that I create a 2D shape and want to attach a 1D lead to only one internal lattice site of a 2D shape. (imagine, that this 1D chain is pointing in the third direction, although this embedding does not matter for physics) Does this make sense? There is a technical problem that 1D families differ from 2D ones and I get a message: "ValueError: Sites with site families (kwant.lattice.Monatomic([[1.0]], [0.0], ''),) do not appear in the system, hence the system does not interrupt the lead." <mailto:kwant-discuss@kwant-project.org> Any hints? Thank you, Sergey

3 8