Trying out mode solving with negative epsilon

[mdecea]

Hello,

I wanted to try out using MPB with negative epsilon materials. I know it is an "experimental" feature, but I have not even been successful in making the MPB simulation complete, I always get the error RuntimeError: meep: invalid dielectric function for MPB. This happens in the "Initializing epsilon function" step.

I have set is_negative_epsilon_ok = True and run ModeSolver.allow_negative_epsilon() before ModeSolver.run(). Am I missing some additional setting to make it work?

Here is a minimal example (in python) reproducing the issue:

import meep as mp
from meep import mpb

geometry_lattice = mp.Lattice(size=mp.Vector3(0,2,2))

mat1 = mp.Medium(epsilon=1)
mat2 = mp.Medium(epsilon=-2)

geometry = [mp.Block(size=mp.Vector3(mp.inf, 2, 1), 
   center=mp.Vector3(z=-0.5), material = mat1), 
   mp.Block(size=mp.Vector3(mp.inf, 2, 1), 
   center=mp.Vector3(z=0.5), material = mat2)

k_min = 0.1
k_max = 2.0
k_points = mp.interpolate(num_k, [mp.Vector3(k_min), mp.Vector3(k_max)])

ms = mpb.ModeSolver(geometry_lattice = geometry_lattice, 
   geometry=geometry, 
   k_points=k_points, 
   resolution = 32, 
   num_bands=2, 
   is_negative_epsilon_ok=True, 
   target_freq=1/mp.inf)

ms.allow_negative_epsilon()
ms.run()

[stevengj]

It's possible that it isn't working in the Python interface, where it may not have been tested. Have you tried it in the Scheme interface?