I tried to use a common 900 W microwave oven melt some alloy with low melting point (138 deg C, eutectic bismut/tin).
The metal did not heat up, neither did the metal container, which makes me think the waves got reflected from it just like on the device inner walls.
The oven is working at 2.45 GHz, about 120 mm wavelength.
I expect it is possible to build some kind of antenna that would heat up?
What would be good shapes - thin and 120 mm long? Half and quater wavelength could work too?
Can the individual long parts be connected - like at the center?
There are microwave kilns available commercially. It's a heat resistant foam material container with an inner graphite-containing coating. I do not know whether they work based on a specific geometry of the coating.