With regards to the spontaneous fire casting bits, in theory metamagic should work as a guide for spell research. Going by at least one reading of the spell research rules, if a spell already exists but isn't on your list, you can just straight up add it to your list for cheap, whereas inventing something whole cloth requires a lot more time and effort.
With that in mind, being able to scribe a scroll of a metamagically enhanced spell should justify scribing that 'version' of the spell into your spellbook.
So for example, learning a [fire] descriptor version of Cone of Cold should only cost 900 gp, and takes one day of research, provided you know the energy substitution metamagic feat, and know Cone of Cold. The cost would go up if you don't know Cone of Cold, or need to hire a spellbook based arcane caster with the relevant feat- or both.
Using such an option would actually introduce a very interesting aspect to the game from a DM perspective, namely 'rare spells'. Metamagic cost reducers exist, after all, so it's entirely feasible that a properly equipped wizard might research a maximized, empowered, widden, quickened fireball with metamagic reduction out the ass (such as via arcane thesis) that takes up only a slightly higher spell slot.
That could then exist as a 'rare spell' that wizard players might quest for, or could be found as loot (in the BBEG's spellbook, for example). It would also give wizards incentive to dump their money on shit other than just more magic items.