Because it is using power from solar panels to accelerate the ions, that's what makes it not a perpetual motion machine. We know we can build ion thrusters that generate enough thrust to overcome the drag at these orbits. But that is with different propellant. The question is if you can do the same using air as the propellant, and with even higher thrust to account for the additional drag a scoop would have over an aerodynamic design.

I suppose solar powered aircraft count as perpetual motion machines in OP's books. They categorically aren't. :)

https://en.wikipedia.org/wiki/Solar-powered_aircraft

The air molecules used by the ion engine as propellant go through channels in the satellite. The molecules aren't "caught" but instead pass through it and once they're ionized, the electromagnetic repulsion between the propellant and the satellite pushes them apart.

As long as the solar panels can supply enough energy, I don't see a problem. It isn't actually a perpetual motion machine as there's a constant energy input and there's no violation of conservation of momentum that I can see. The problem I think is more of an optimization - how big can they make the solar panels without adding too much drag? That will set a bound for how low these things can fly.

Universe Today has a great interview on this tech. Even optimistically, it's knifes edge, just barely enough margin to keep things going before factoring in things like solar storms.

Why would friction always be higher? In principle, this is just like an electric RC plane. You use the surrounding air for propulsion while keeping your own mass constant.

Any sort of energy harvesting and reuse is going to involve some sort of losses along the way. Plus not all surfaces of the craft will be energy harvesting so those parts are just straight air friction.

The energy for the propulsion is coming from solar power.