Satellites comprised mainly of aluminium burn up in the upper atmosphere. This likely reduces the ozone layer, as that's where the chemical reactions take place, unlike industrial activity on the ground. A certain amount of aluminium gets injected into the upper atmosphere by meteorites without any human activity involved, but the quantities burned up by modern satellite constellations will significantly exceed that. Nothing different about Starlink satellites compared with anyone else's operating in LEO, but they do operate more than anyone else.
There's considerably more aluminum oxides in the stratosphere from solid rocket motors than satellite reentires, and that's been going on since the 1970's. The 135 Space Shuttle launches burned 160 tons of aluminum powder, per launch, in their SRB's. (22 kilotons in total?) An Ariane 6 launch burns 55 or 110 tons (2 or 4 boosters).
The largest satellite payloads are 20 tons in a launch.
(Al2O3 isn't even the highest-impact ozone depletor within the space industry; that's chlorine. Also from solid rocket fuel).
Yes, I agree launches also contribute and increase frequency is also a plausible risk factor (and is significant increases in launch frequency themselves driven by megaconstellations, though I don't think the Falcon 9 is as problematic as the Shuttle was) Main difference is that the satellites lose most of their mass in the relevant zone, whereas launches pass through it as quickly as possible. Satellites are small, but Starlink's deorbiting cycle is expected to reach 2 tonnes per day.