I am using COMSOL v5.3a to simulate an infinite interface between an air dielectric and a dispersive medium using a periodic boundary along x in 2D. I defined a rectangle, with a horizontal centerline on the x-axis acting as an interface that separates two symmetric y > 0 and y < 0 domains. I define two wave equations for the two materials. Wave Equation, Electric 1 uses permittivity of air above the interface where y>0. Wave Equation, Electric 2 is a negative permittivity material from a Drude model where y < 0. The oscillator strength is 1, the resonance frequency 0, and the damping time is 0. The left and right vertical walls are periodic boundaries with a Floquet periodicity of 2k0 (twice the wavenumber of free-space). I use the eigenfrequency study and let comsol choose the solver PARADISO usually. I successfully had achieved surface plasmon polaritons along the interface with a frequency-independent material, but when I use the Drude model, even though it has slight negative permittivity ~ -1, there is no radiation inside or at the boundary in either an eigenfrequency or frequency domain study. I select the plasma frequency to be 2 pi sqrt(2) multiplied by 10 GHz, and I search for eigenfrequencies at 10 GHz.
If anyone has successfully used this a dispersive medium for this purpose in the frequency domain, I would appreciate any assistance that can be offered. Thank-you for reading. For more information, please see the attached mph file.