Hi @conpierce8 , thank you very much for your help! I see I did a rookie mistake with the ds, classic. By correcting the ds definition the imaginary part of the solution matches that of Elmer (qualitatively, the values are off). The real part is still zero.
If, on top of that, I replace my bilinear and linear forms with yours I still get a wrong solution: identical real and imaginary parts. Maybe that’s the right direction though. A bilinear form for this problem was provided here. I am not sure whether that is correct, but I guess I will try to make sure I am implementing that and see how it goes.