Dear MPB users, I found a weird output of the epsilon tensor in MPB while trying to simulate an anisotropic nanofiber. Here, the epsilon.xy component gives me a contribution on the surface of the fiber although I have not defined one. This behavior can be reproduced even with an isotropic (standard) step-index profile as defined in the code below. Just copy the code to a file, run it, and then compare the outputs of the following commands:
h5topng epsilon.h5:epsilon_inverse.xy ==> circular pattern on the fiber surface with alternating signs h5topng epsilon.h5:epsilon_inverse.yz h5topng epsilon.h5:epsilon_inverse.xz ==> 0 in the whole region (black) I have cross-checked the h5-file with Mathematica and find matching results: 4 pairs of two spikes with a magnitude of 0.1 in the diagonals with alternating signs. Has anyone an idea whether this is real or just in the output function? Every help is appreciated. Regards, Christian Version: MPB 1.4.2, using libctl 3.1 and Guile 1.8.8 (=current Ubuntu) Code: ===== ;Fiber parameters (define-param mu 0.5); Diameter (micrometer) (define (eps-func p) (make dielectric (epsilon (* 1.45258 1.45258)) ) ) (define clad (make material-function (material-func eps-func))) (define-param radclad 0.5) (define-param wl .850); wavelength (um) (define-param omega (/ mu wl)) ; conversion to mpb-units (define-param sc-x 5) ; supercell width (mu) (define-param sc-y 5) ; supercell height (mu) (set! geometry-lattice (make lattice (size sc-x sc-y no-size))) ; define the fiber geometry (set! geometry (list (make cylinder (center 0 0 0) (height infinity) (axis 0 0 1) (material clad) (radius radclad)) ) ) (set-param! num-bands 2) (set-param! resolution 100) ; pixels/um (define-param bestguess 1.14504) ; Output the x component of the Poynting vector for num-bands bands at omega (find-k NO-PARITY omega 1 num-bands (vector3 0 0 1) 1e-5 (* omega bestguess) (* omega .5) (* omega 2) fix-efield-phase output-efield display-group-velocities ) _______________________________________________ mpb-discuss mailing list mpb-discuss@ab-initio.mit.edu http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/mpb-discuss