Discussion Forum

Hi,

I am trying to simulate in time domain the propagating wave along a metal/dielectric interface, where the metal permittivity is following the drude model. The way I approached that problem is that I added PDE mode to the model solving the following equation for the polarization P

d^2/dt^2 (P)=Epsilon0*(Wp^2)*E(t)

Then I input this P into the metal domains, instead of using some value of epsilon as done conventionally

where Epsilon0 is the free space permittivity, Wp is the angular plasma frequency, and E(t) is the electric field.


When I use Wp that is much lower than my operating frequency, i.e the metal behaves as a positive epsilon dielectric, the setup works very nicely, however if i just raise Wp, so that I have negative epsilon medium (which is what I need to simulate), the model gives results that does not make any sense. I tried increasing the mesh at the interface and tried many other tricks that I thought may be the reason but still not working, any suggestions ?

Thank you so much !!