Um, yeah, I was inaccurate there. It’s not a true chemical bond of aluminum to steel. But is there not some adhesion of the corrosion products to both substrates? I can think of few cases of corrosion I’ve seen, where removal to “bare metal” does not involve force / scraping / sanding. It usually doesn’t just rub off to bare metal with one’s (unaided) finger...
There is also a frictional effect too, no? That is, the surfaces become pitted, so one has to shear the corrosion products to break free. It would be somewhat similar (but not exactly so) to the effect of fine grit in the threads.
My other question is, does the penetrant actually break down the corrosion products? I would think not, but I could be wrong.
There may be a wedging effect and seizure taking place from the tapered bolt being squeezed/tightened down into the softer aluminum seat, something like the seizing effect that occurs on a tie rod end getting stuck in its seat. Modest heat applied to the aluminum around the bolt should address this issue.
Some of the corrosion products usually adhere pretty tightly to the base metal: i.e. Aluminum Oxide to Aluminum, and Iron Oxide to Iron. They do not cross-adhere.
Increased volume is the major effect, but the surfaces become less smooth and there is an increase in friction.
The third question is an ambiguous one. The penetrant does not chemically break down the corrosion products, but it does promote a physical breakdown.
With steel in contact to aluminum, the aluminum will preferentially corrode, acting as a sacrificial anode for the steel.