This is a little interesting, but I don't know if it would work as simply (if difficultly) as that, especially if the iron winds up bound in complex chemical arrangements with other things; I mean, you're less likely to get hematite out of that as you are <insert just about any volcanic stone containing iron as well as a bunch of other things in a complex arrangement>.
"Less likely", in a sense, doesn't need to be a real issue - if we were to make it into a reaction, we could make it into a 5% chance of getting an ore with enough iron out of the reaction to be smeltable (like hemitite), and it would still be useful. Throw enough dice and eventually, the law of averages will favor you.
I mostly think about it because the "gameplay entropy" doesn't really sit well with me - you start with a mountain, and then subtract out a finite supply of resources from it. Eventually, you can just plain run out of EARTH, which is pretty silly.
I've been kicking around an idea to have some sort of vulcanism mechanic that would be able to trigger shifts in vulcanism when a certain critical threshold is hit that makes the magma sea simply vomit up some magma which eventually cools into semi-molten, then regular rock to replinish the earth when you actually manage to pretty much excavate the entire crust of the planet down to the mantle. When it does that, it becomes not just obsidian, but also things like gabbro and basalt, and relevant ores might be found there, as well.
Basically, I view the world as an ecosystem, where matter is continuously in flux, but not being created or destroyed. This entropic decay of the game that forces regenning worlds and abandoning forts and player-made towns being incapable of living more than about 20 years or so really bugs me at a fundamental level.
If stone is getting wiped out of the game by being dumped into the magma sea to eliminate the excess useless stone, then why not have some stone capable of coming back out of the magma sea?