Discussion Forum

A "point dipole" radiator and a "dipole antenna" are related, but are not exactly the same thing. See, for example, the discussions at: https://en.wikipedia.org/wiki/Dipole_antenna#Short_dipole and https://en.wikipedia.org/wiki/Dipole_antenna#Hertzian

What kind of dipole do you mean?

You are right. I mean a 'point dipole' radiatior, the one defined as Hertzian in the link above, with near field non radiating component and far-field. The field emitted by a finite oscillating current density (up to first order) as derived in Novotny and Hecht: https://www.cambridge.org/core/books/principles-of-nanooptics/EAA3E0D04179A7252088E27483C4ACC1

Well, among other approaches, you can create a small narrow cylinder (a vertical rectangle of length L<< lambda and width R<<L, in 2D axisym) subtract it from a (much larger) computational space, and then impose a strictly-uniform (constant) z-directed current along the rectangle's vertical boundary, as a surface current-density boundary condition. If this rectangle's dimensions are sufficiently small compared to any other items (or distances from the dipole) of interest in your model, then it will probably work well enough as a Hertzian dipole. Alternatively, you could employ the Scattered Field formulation with the Hertzian dipole fields analytically specified by you (since, after all, these fields are known quantities), and entered as "user-defined". Those are two that come to my mind, but there may be other possible and practical options.

Thanks! I will work on the frist option since I will play with the environment and design the material and shape in the vicinity of the dipole. Thus, the scattered fields will not be known in principle.