Interesting! -- Your observations are certainly consistent with my experience (assuming the 'altered' diodes' reverse leakage indeed owes to 'junction insufficiency' -- as opposed to external 'electrolytic conduction') -- Though I must confess to no small surprise! - Apparently junction fabrication in rectifier diodes varies significantly from that in BJTs? -- I have, for instance, seen several (mass produced) designs wherein the B-E junctions of Si BJTs are biased for Zener operationAs the electrons are swept across the broken down junction by 1400 V they can pick up enough energy to damage the lattice since the junction was not designed to operate that way (and, in all likelihood, it's ability to behave well under those conditions was sacrificed in order to get better performance under the conditions that it WAS designed to operate under. I would not be surprised that the sacrifices made to achieve the tradeoff are more severe for diodes that are designed to carry higher and higher forward currents.
Many thanks and very best regards
HP