Dear all, I want to plot the local density of states (LDOS) for a honeycomb system of a rectangular shape modeled by Hamiltonian H. Because the Bravais lattice vectors are non-orthogonal, the honeycomb leads (but modeled by different Hamiltonian H') are attached in such a way, that sites have to be added to the scattering region. This added part, however, is modeled by H' whereas the rest (original scattering region) by H. I know that I can just say that the added part is actually part of the leads (matter of definition). But, if I want to plot LDOS using "kwant.ldos(sys, energy)", it plots LDOS also on this added part modeled by H', which obscures the LDOS of my interest on the original scattering region modeled by H. How to make KWANT not to plot the LDOS on the artificially added part to the scattering region? Thanks for an advice, Tibor
Hi Tibor, you wrote:
But, if I want to plot LDOS using "kwant.ldos(sys, energy)", it plots LDOS also on this added part modeled by H', which obscures the LDOS of my interest on the original scattering region modeled by H.
The function kwant.plotter.map() does not provide a way to exclude certain sites from being plotted. I see several possible solutions to your problem: • Use a somewhat obscure trick [1] to avoid adding lead sites to the scattering region. (This is the easiest approach, but it may not be possible if your scattering region does not have a linear interface.) • Take the code of kwant.plotter.map() and modify it such that certain sites can be hidden. Then, save the sites of the system before the leads are added and pass them to the modified map(). • Plot the data in some other home-made way. Let us know if any of the above works for you. Cheers, Christoph [1] http://thread.gmane.org/gmane.comp.science.kwant.user/170/focus=175
participants (2)
-
Christoph Groth -
Tibor Sekera