Steven G. Johnson <[email protected]> wrote: > On Aug 14, 2009, at 12:54 PM, [email protected] wrote: > > I'd like to simulate magneto-electric materials (etc), > > i.e. where D depends on H, and B on E. > > As far as I can tell, meep doesn't do this. Would it > > be hard to implement?
> Meep doesn't currently support this. > The tricky part about implementing such a thing is that the D/E and B/ > H fields in FDTD are stored at different times (offset by half a > timestep), so if you want second-order accuracy you need to average > over a couple of timesteps to couple the two fields in this way. > (This is do-able; the easiest way is probaby to save the backup fields > in a copy of the array, which sacrifices memory but avoids having to > modify the curl equations etc....this is important because you want to > avoid a combinatorial explosion of the number of possible cases in the > curl equations.) > Do you need a dispersive version of this, or is nondispersive enough? > [...] I'd like dispersive, but I'll settle for non-dispersive if I have to. And I think you could do non-dispersive in a way that allows simulation of the Minkowski constitutive relations for moving media (see McCall & Censor AJP75, 1134 (2007) doi 10.1119/1.2772281, which would be nice. > How common are real materials that have these properties? (We mostly > work at infrared frequencies in our group, so we don't come across > magnetic materials often.) They're not that uncommon -- but, more to the point, they're interesting. -- ---------------------------------+--------------------------------- Dr. Paul Kinsler Blackett Laboratory (Photonics) (ph) +44-20-759-47734 (fax) 47714 Imperial College London, [email protected] SW7 2AZ, United Kingdom. http://www.qols.ph.ic.ac.uk/~kinsle/ _______________________________________________ meep-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
