Epoxy is a rather inert polymer. Solvents or normal acids would take very long to remove it. Being an oxidizer, nitric acid does it faster, and heat accelerates it.
Said "epoxy" is filled with a powder of quite pure silica in huge proportion, supposedly by mixing different grain sizes. This stabilizes the dimensions against humidity and temperature to match silicon better. After etching, the silica remains in the acid and thickens it.
This epoxy is very special in that its polymerization or cross-linking releases no HCl - not even that little bit, because HCl catalyses the corrosion by humidity of the ultra-thin metal layers. Long ago, the only plant worldwide burned in Japan and the whole semiconductor industry was in trouble.
There may well be additional passivation layers between the chip and the epoxy, like polyimide, and these layers can react differently to the acids, leaving black residues. Just a hypothesis.
If your chips still use bonding wires, these are of gold (aluminium existed long ago). Untouched by both acids, so the recovery would rather be mechanical. If TAB makes the connections between the chip and the macroscopic contacts (as was the case in chipcards), it usually had gold conductors printed on polyimide.
In case the chips still have aluminium conductors, anything exposed goes bye-bye in the acids. But under the silica or silicon nitride passivation, it can survive.
Please take with mistrust: I left the semiconductors three decades ago, when we made the etchings with flint stone and the ion implantation with a blowpipe.