The most extreme "conventional" chemical reactions are generally those between the elements at opposite ends of the periodic table (ignoring the "noble gasses"). Sodium reacting with Fluorine is supposedly one of the most volatile combinations possible, with Calcium and Chlorine being one step down from each of those, at least as far as I understand from my non-specialist viewpoint. Chemicals of that volatility have their own unique storage issues, which is part of why they aren't used in that manner today; another part being that the energy release rate is so rapid that containing/channeling the explosion is far more difficult than with conventional explosives.
Energy alone is not a determining factor in propellants, be they rocket fuel or gunpowder. An additional factor of considerable import is the reaction products. Hydrogen-oxygen combustion is great for rockets because the only products are gases (unless you severely over-expand the nozzle and start getting water and ice condensation). Kerosene-oxygen is okay, but its gaseous products are heavier and thus have a lower exhaust velocity. And, if you run fuel-rich (to keep wall temperatures down) you'll have solids - soot - in the exhaust, too, which is a problem in reusable rockets.
Speaking of alkaline-halogen reactants, lithium-fluoride rockets have been considered for use, but their product (lithium fluoride) wasn't a well-behaved gas; it only boils at 3000F, so you'll see a lot of droplets in the exhaust stream (which is hard on everything). Further, to make the most of it, you had to mix in hydrogen. The resulting rocket was an engineering nightmare: tankage for low density ultra-cold hydrogen, low density hot molten lithium, and toxic, cold liquid fluorine combined in a plumbing horror of an engine. All the sacrifices made cancelled the substantial gain in specific impulse compared to hydrogen-oxygen. Sodium-fluorine might also release a lot of energy, but the result will more of a caustic wet burp than an energetic propulsive gas explosion.
Aluminum-oxygen is also highly energetic and both elements are available plentifully on the moon (if you need in-situ production of rocket fuel), but the products of the reaction are abrasive super-hard sapphire powders and droplets with atrocious behavior in a rocket nozzle.
Recent 20th century "double base" powders cause a compound set of reactions, generating more overall energy than a simple single reaction,
Modestly, yes. The nitroglycerin does help the nitrocellulose. As with the better rocket fuels noted above, the ideal products of modern nitro-based gunwpowders are mostly gases.
I won't entirely give up on chemical reactions for ACs. Explosives do expand at up to 7-8km/s (far faster than real “gunpowdersâ€), which is enough for short-ranged ACs in space, but explosives are generally poor projectile propellants, hence the move to nuclear detonations in NACs. Use brute force and ablation to achieve what chemical impulses can’t.
The introduction of depleted uranium or other "heavy metals" as penetrators makes it possible to make use of those higher muzzle velocities. Previous hardened steel penetrators would shatter on impact at those velocities, paradoxically doing less damage than if travelling slightly slower.
Depleted uranium is not a particularly strong material, either; high-end projectile steels are universally harder and stronger than DU. Uranium’s two advantages over steel are density (so more kinetic energy is applied to less armor) and a self-sharpening shearing behavior as it plows through armor. Tungsten, which is almost equally as dense as DU and much, much stronger than uranium, lacks the self-sharpening behavior and generally sees 10-30% less penetration because of it.
Speaking generally, as impact velocities exceed ~2km/s, there isn’t time for a projectile to shatter. Impacts are modeled with hydrocode software because armor and projectile both behave like thick liquids because there isn’t time for fractures to spread before the impact is over.