see these academic papers. XPS is a great analytical instrument but prohibitively expensive bit of kit
It and other techniques likely point to silver sulfide, but can’t rule out copper oxide.
The chemistry is similar to that well known for removing tarnish from silver coins, since the tarnish is likely Ag2S in both cases looking at the above research papers (even if it’s copper the same process works) a similar treatment should and did work.
To find out why the chemistry behind the silver tarnish removal (and similar process here) I point you to redox potentials of the reactions:
Cu2O(s)+H2O+2e- <—> 2Cu(s)+OH- -0.36
Al3+ + 3e- <—> Al(s) -1.662
Ag2S(s) + 2e- <—> 2Ag(s)+ S2-(aq) -0.69
The more negative the potential the more the equilibrium lies to the left. Since Al can easily donate electrons to either silver sulfide or copper oxide it can reduce either. Another forum member did this recently on his libertad and worked, the above is the reason why.